Skip to contentSkip to conversation

Reverse the order of OVERRIDES

PATCH In review Chris Larson 11 participants 36 messages Oct 15, 2010 – Dec 1, 2010

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Reverse the order of OVERRIDES

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that. As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific. As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so clean those up as well.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally, and it's then translated to : separated when used in OVERRIDES.

---
 conf/bitbake.conf                             |   17 +++--------------
 conf/distro/include/arm-thumb.inc             |    6 +++---
 conf/distro/micro.conf                        |    6 +++---
 conf/distro/minimal.conf                      |    5 +++--
 conf/distro/shr.conf                          |    5 +++--
 conf/machine/fsg3be.conf                      |    2 +-
 conf/machine/kixrp435.conf                    |    2 +-
 conf/machine/nslu2be.conf                     |    2 +-
 conf/machine/nslu2le.conf                     |    2 +-
 recipes/binutils/binutils_csl-arm-20050416.bb |    2 +-
 recipes/binutils/binutils_csl-arm-20050603.bb |    2 +-
 recipes/eglibc/eglibc-package.bbclass         |    2 +-
 recipes/glibc/glibc-package.inc               |    2 +-
 recipes/prelink/prelink_20061027.bb           |    2 +-
 recipes/prelink/prelink_20071009.bb           |    2 +-
 recipes/prelink/prelink_20090925.bb           |    2 +-
 16 files changed, 26 insertions(+), 35 deletions(-)
conf/bitbake.conf +3 -14
index 494e14e..c2359dc 100644
@@ -675,20 +675,9 @@ BUILDCFG_NEEDEDVARS ?= "TARGET_ARCH TARGET_OS"
 
 # Overrides are processed left to right, so the ones that are named later take precedence.
 # You generally want them to go from least to most specific.
-# 
-# This means that an envionment variable named '<foo>_arm' overrides an
-# environment variable '<foo>' (when ${TARGET_ARCH} is arm).
-# an environment variable '<foo>_ramses' overrides '<foo>' but doesn't override
-# '<foo>_arm' when ${MACHINE} is 'ramses'. 
-# If you use combination ie '<foo>_arm_ramses', then '<foo>_arm_ramses' will override 
-# '<foo>_arm' and then '<foo>' will be overriden with that value from '<foo>_arm'.
-# And finally '<foo>_local' overrides anything, but with lowest priority.
-#
-# This works for  functions as well, they are really just environment variables.
-# Default OVERRIDES to make compilation fail fast in case of build system misconfiguration.
-OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
-# Alternative OVERRIDES definition without "fail fast", usually only for native building and Scratchbox toolchains.
-#OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:pn-${PN}"
+MACHINE_OVERRIDES ?= ""
+OVERRIDES = "pn-${PN}:fail-fast:build-${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:\
+${DISTRO}:${@':'.join(d.getVar('MACHINE_OVERRIDES', True).split())}:${MACHINE}:local"
 
 ##################################################################
 # Include the rest of the config files.
conf/distro/include/arm-thumb.inc +3 -3
index 36edd4f..75ed978 100644
@@ -16,9 +16,9 @@ THUMB_INTERWORK ?= "no"
 #    arm system and vice versa.  It is strongly recommended that DISTROs not
 #    turn this off - the actual cost is very small.
 
-OVERRIDE_THUMB = "${@['', ':thumb'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
-OVERRIDE_INTERWORK = "${@['', ':thumb-interwork'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
-OVERRIDES_append_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
+OVERRIDE_THUMB = "${@['', 'thumb:'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
+OVERRIDE_INTERWORK = "${@['', 'thumb-interwork:'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
+OVERRIDES_prepend_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
 
 #    Compiler and linker options for application code and kernel code.  These
 #    options ensure that the compiler has the correct settings for the selected
conf/distro/micro.conf +3 -3
index 6d396b3..19b5e62 100644
@@ -69,10 +69,10 @@ require conf/distro/include/sane-toolchain.inc
 require conf/distro/include/arm-thumb.inc
 
 #############################################################################
-# OVERRIDES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:\
-${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # PREFERRED VERSIONS
conf/distro/minimal.conf +3 -2
index 376ffde..0d32cfe 100644
@@ -81,9 +81,10 @@ KERNEL = "kernel26"
 MACHINE_KERNEL_VERSION = "2.6"
 
 #############################################################################
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # TOOLCHAIN
conf/distro/shr.conf +3 -2
index ade8cba..74e3fe9 100644
@@ -151,9 +151,10 @@ KERNEL = "kernel26"
 MACHINE_KERNEL_VERSION = "2.6"
 
 #############################################################################
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # TOOLCHAIN
conf/machine/fsg3be.conf +1 -1
index a2ca95b..c5ddc25 100644
@@ -8,7 +8,7 @@ PACKAGE_EXTRA_ARCHS = "ixp4xxbe"
 
 MACHINE_ARCH = "ixp4xxbe"
 
-OVERRIDES = "local:${MACHINE}:ixp4xxbe:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx ixp4xxbe"
 
 # Match the vendor's latest kernel
 PREFERRED_PROVIDER_virtual/kernel ?= "fsg3-kernel"
conf/machine/kixrp435.conf +1 -1
index 8259127..b4a2fa6 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"
 
 PREFERRED_PROVIDER_virtual/kernel = "linux"
 PREFERRED_VERSION_linux = "2.6.30"
conf/machine/nslu2be.conf +1 -1
index 9bc92f1..88e6ed2 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "armeb"
 # Note: armv4b armv4tb armv5teb will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5eb ixp4xxbe"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"
 
 ROOT_FLASH_SIZE ?= "6"
 
conf/machine/nslu2le.conf +1 -1
index 62e47cb..edd0d94 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"
 
 ROOT_FLASH_SIZE ?= "6"
 
recipes/binutils/binutils_csl-arm-20050416.bb +1 -1
index ad45438..71803d1 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"
 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs"
 PV = "2.15.99+csl-arm+cvs20050416"
 PR = "r3"
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 DEFAULT_PREFERENCE = "-1"
 DEFAULT_PREFERENCE_arm-linux = "-1"
 DEFAULT_PREFERENCE_arm-linuxeabi = "-1"
recipes/binutils/binutils_csl-arm-20050603.bb +1 -1
index 595cb98..50e8677 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"
 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs"
 PV = "2.15.99+csl-arm+cvs20050603"
 PR = "r3"
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 DEFAULT_PREFERENCE = "-1"
 #DEFAULT_PREFERENCE_arm-linux = "1"
 #DEFAULT_PREFERENCE_arm-linuxeabi = "1"
recipes/eglibc/eglibc-package.bbclass +1 -1
index dbca1ab..c802bd5 100644
@@ -116,7 +116,7 @@ def get_eglibc_fpu_setting(bb, d):
 EXTRA_OECONF += "${@get_eglibc_fpu_setting(bb, d)}"
 EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"
 
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 
 do_configure_prepend() {
         sed -e "s#@BASH@#/bin/sh#" -i ${S}/elf/ldd.bash.in
recipes/glibc/glibc-package.inc +1 -1
index 0b263bb..8cd8564 100644
@@ -44,7 +44,7 @@ def get_glibc_fpu_setting(bb, d):
 EXTRA_OECONF += "${@get_glibc_fpu_setting(bb, d)}"
 EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"
 
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 
 do_install() {
 	oe_runmake install_root=${D} install
recipes/prelink/prelink_20061027.bb +1 -1
index ad2faf5..552117b 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 TARGET_OS_ORIG := "${TARGET_OS}"
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"
recipes/prelink/prelink_20071009.bb +1 -1
index faf930c..e2ab3fb 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 TARGET_OS_ORIG := "${TARGET_OS}"
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"
recipes/prelink/prelink_20090925.bb +1 -1
index 4c83402..5b247b6 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 #TARGET_OS_ORIG := "${TARGET_OS}"
-#OVERRIDES_append = ":${TARGET_OS_ORIG}"
+#OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 #SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"
Signature · 1 line
1.7.2.3
Acked-by on v1 1/1
#1 lore raw
Quoted 13 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that.  As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific.  As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so clean those up as well.

Quoted 8 lines

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go.  This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally, and it's then translated to : separated when used in OVERRIDES.

both things look nice to me. We might need to document MACHINE_OVERRIDES though.

Quoted 268 lines

---
 conf/bitbake.conf                             |   17 +++-------------- conf/distro/include/arm-thumb.inc             |    6 +++--- conf/distro/micro.conf                        |    6 +++--- conf/distro/minimal.conf                      |    5 +++-- conf/distro/shr.conf                          |    5 +++-- conf/machine/fsg3be.conf                      |    2 +-

 conf/machine/kixrp435.conf                    |    2 +-
 conf/machine/nslu2be.conf                     |    2 +-
 conf/machine/nslu2le.conf                     |    2 +-

 recipes/binutils/binutils_csl-arm-20050416.bb |    2 +- recipes/binutils/binutils_csl-arm-20050603.bb |    2 +- recipes/eglibc/eglibc-package.bbclass         |    2 +- recipes/glibc/glibc-package.inc               |    2 +-

 recipes/prelink/prelink_20061027.bb           |    2 +-
 recipes/prelink/prelink_20071009.bb           |    2 +-
 recipes/prelink/prelink_20090925.bb           |    2 +-
 16 files changed, 26 insertions(+), 35 deletions(-)
conf/bitbake.conf +3 -14
index 494e14e..c2359dc 100644
@@ -675,20 +675,9 @@ BUILDCFG_NEEDEDVARS ?= "TARGET_ARCH TARGET_OS"
 
  # Overrides are processed left to right, so the ones that are named later take precedence.
  # You generally want them to go from least to most specific.
-#
-# This means that an envionment variable named '<foo>_arm' overrides an
-# environment variable '<foo>' (when ${TARGET_ARCH} is arm).
-# an environment variable '<foo>_ramses' overrides '<foo>' but doesn't override
-# '<foo>_arm' when ${MACHINE} is 'ramses'.
-# If you use combination ie '<foo>_arm_ramses', then '<foo>_arm_ramses' will override
-# '<foo>_arm' and then '<foo>' will be overriden with that value from '<foo>_arm'.
-# And finally '<foo>_local' overrides anything, but with lowest priority.
-#
-# This works for  functions as well, they are really just environment variables.
-# Default OVERRIDES to make compilation fail fast in case of build system misconfiguration.
-OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
-# Alternative OVERRIDES definition without "fail fast", usually only for native building and Scratchbox toolchains.
-#OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:pn-${PN}"
+MACHINE_OVERRIDES ?= ""
+OVERRIDES = "pn-${PN}:fail-fast:build-${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:\
+${DISTRO}:${@':'.join(d.getVar('MACHINE_OVERRIDES', True).split())}:${MACHINE}:local"
 
  ##################################################################
  # Include the rest of the config files.
conf/distro/include/arm-thumb.inc +3 -3
index 36edd4f..75ed978 100644
@@ -16,9 +16,9 @@ THUMB_INTERWORK ?= "no"
  #    arm system and vice versa.  It is strongly recommended that DISTROs not
  #    turn this off - the actual cost is very small.
 
-OVERRIDE_THUMB = "${@['', ':thumb'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
-OVERRIDE_INTERWORK = "${@['', ':thumb-interwork'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
-OVERRIDES_append_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
+OVERRIDE_THUMB = "${@['', 'thumb:'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
+OVERRIDE_INTERWORK = "${@['', 'thumb-interwork:'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
+OVERRIDES_prepend_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
 
  #    Compiler and linker options for application code and kernel code.  These
  #    options ensure that the compiler has the correct settings for the selected
conf/distro/micro.conf +0 -0
index 6d396b3..19b5e62 100644
@@ -69,10 +69,10 @@ require conf/distro/include/sane-toolchain.inc

 require conf/distro/include/arm-thumb.inc

 #############################################################################

conf/distro/micro.conf +1 -1
-# OVERRIDES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES

 #############################################################################

conf/distro/micro.conf +2 -2
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:\
-${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

 #############################################################################
 # PREFERRED VERSIONS

conf/distro/minimal.conf +0 -0
index 376ffde..0d32cfe 100644
@@ -81,9 +81,10 @@ KERNEL = "kernel26"

 MACHINE_KERNEL_VERSION = "2.6"

 #############################################################################

conf/distro/minimal.conf +1 -1
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES

 #############################################################################

conf/distro/minimal.conf +2 -1
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

 #############################################################################
 # TOOLCHAIN

conf/distro/shr.conf +0 -0
index ade8cba..74e3fe9 100644
@@ -151,9 +151,10 @@ KERNEL = "kernel26"

 MACHINE_KERNEL_VERSION = "2.6"

 #############################################################################

conf/distro/shr.conf +1 -1
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES

 #############################################################################

conf/distro/shr.conf +2 -1
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

 #############################################################################
 # TOOLCHAIN

conf/machine/fsg3be.conf +0 -0
index a2ca95b..c5ddc25 100644
@@ -8,7 +8,7 @@ PACKAGE_EXTRA_ARCHS = "ixp4xxbe"
 

 MACHINE_ARCH = "ixp4xxbe"

conf/machine/fsg3be.conf +1 -1
-OVERRIDES = "local:${MACHINE}:ixp4xxbe:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx ixp4xxbe"

 # Match the vendor's latest kernel PREFERRED_PROVIDER_virtual/kernel ?= "fsg3-kernel"

conf/machine/kixrp435.conf +0 -0
index 8259127..b4a2fa6 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
  # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.

 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

conf/machine/kixrp435.conf +1 -1
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"

 PREFERRED_PROVIDER_virtual/kernel = "linux"
 PREFERRED_VERSION_linux = "2.6.30"

conf/machine/nslu2be.conf +0 -0
index 9bc92f1..88e6ed2 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "armeb"
  # Note: armv4b armv4tb armv5teb will be added by tune-xscale.inc automatically.

 PACKAGE_EXTRA_ARCHS = "armv5eb ixp4xxbe"

conf/machine/nslu2be.conf +1 -1
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"

 ROOT_FLASH_SIZE ?= "6"

conf/machine/nslu2le.conf +0 -0
index 62e47cb..edd0d94 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
  # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.

 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

conf/machine/nslu2le.conf +1 -1
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"

 ROOT_FLASH_SIZE ?= "6"

recipes/binutils/binutils_csl-arm-20050416.bb +0 -0
index ad45438..71803d1 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"

 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs" PV = "2.15.99+csl-arm+cvs20050416"
 PR = "r3"

recipes/binutils/binutils_csl-arm-20050416.bb +1 -1
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"

 DEFAULT_PREFERENCE = "-1"
 DEFAULT_PREFERENCE_arm-linux = "-1"
 DEFAULT_PREFERENCE_arm-linuxeabi = "-1"

recipes/binutils/binutils_csl-arm-20050603.bb +0 -0
index 595cb98..50e8677 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"

 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs" PV = "2.15.99+csl-arm+cvs20050603"
 PR = "r3"

recipes/binutils/binutils_csl-arm-20050603.bb +1 -1
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"

 DEFAULT_PREFERENCE = "-1"
 #DEFAULT_PREFERENCE_arm-linux = "1"
 #DEFAULT_PREFERENCE_arm-linuxeabi = "1"

recipes/eglibc/eglibc-package.bbclass +0 -0
index dbca1ab..c802bd5 100644
@@ -116,7 +116,7 @@ def get_eglibc_fpu_setting(bb, d):

 EXTRA_OECONF += "${@get_eglibc_fpu_setting(bb, d)}" EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"

recipes/eglibc/eglibc-package.bbclass +1 -1
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"

 do_configure_prepend() {
        sed -e "s#@BASH@#/bin/sh#" -i ${S}/elf/ldd.bash.in

recipes/glibc/glibc-package.inc +0 -0
index 0b263bb..8cd8564 100644
@@ -44,7 +44,7 @@ def get_glibc_fpu_setting(bb, d):

 EXTRA_OECONF += "${@get_glibc_fpu_setting(bb, d)}" EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"

recipes/glibc/glibc-package.inc +1 -1
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"

 do_install() {
       oe_runmake install_root=${D} install

recipes/prelink/prelink_20061027.bb +0 -0
index ad2faf5..552117b 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \

           file://prelink.default"

 TARGET_OS_ORIG := "${TARGET_OS}"

recipes/prelink/prelink_20061027.bb +1 -1
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"

 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"

 S = "${WORKDIR}/prelink-0.0.${PV}"

recipes/prelink/prelink_20071009.bb +0 -0
index faf930c..e2ab3fb 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \

           file://prelink.default"

 TARGET_OS_ORIG := "${TARGET_OS}"

recipes/prelink/prelink_20071009.bb +1 -1
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"

 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"

 S = "${WORKDIR}/prelink-0.0.${PV}"

recipes/prelink/prelink_20090925.bb +0 -0
index 4c83402..5b247b6 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \

           file://prelink.default"

 #TARGET_OS_ORIG := "${TARGET_OS}"

recipes/prelink/prelink_20090925.bb +1 -1
-#OVERRIDES_append = ":${TARGET_OS_ORIG}"
+#OVERRIDES_prepend = "${TARGET_OS_ORIG}:"

 #SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"

 S = "${WORKDIR}/prelink-0.0.${PV}"
--
1.7.2.3

_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

#2 lore raw

2010/10/15 Khem Raj <[email protected]>:

Quoted 21 lines · Chris Larson
Quoted 20 lines

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that.  As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific.  As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so clean those up as well.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go.  This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally, and it's then translated to : separated when used in OVERRIDES.

Quoted 6 lines
Quoted 1 line

both things look nice to me. We might need to document MACHINE_OVERRIDES though.

Looks fine to me.
Actually I've wondered several times why local did not have priority (but I had to do things like X_local).
But before adding an ack, I would like to ask what has been done wrt testing (as this is potentially quite invasive).

Also is there a way we can detect where the recipes can be simplified by adding unneeded appendices? (e.g by modifying/extending recipe_sanity).

Thanks for the patch! Frans

#3 lore raw
In reply to Khem Raj
Quoted 21 lines · Khem Raj
Quoted 20 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that. As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific. As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so clean those up as well.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally, and it's then translated to : separated when used in OVERRIDES.

Quoted 3 lines
Quoted 1 line

both things look nice to me. We might need to document MACHINE_OVERRIDES though.

Yes, no new variables without updating the docs.

Signature · 2 lines
Tom Rini
Mentor Graphics Corporation
Acked-by on v1 1/1
#4 lore raw
In reply to Chris Larson

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 13 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that. As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific. As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so clean those up as well.

Quoted 8 lines

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally, and it's then translated to : separated when used in OVERRIDES.

Thanks for the patch! Chase or I will take care of moving SOC_FAMILY to MACHINE_OVERRIDES after this lands.

Quoted 251 lines
---
 conf/bitbake.conf                             |   17 +++--------------
 conf/distro/include/arm-thumb.inc             |    6 +++---
 conf/distro/micro.conf                        |    6 +++---
 conf/distro/minimal.conf                      |    5 +++--
 conf/distro/shr.conf                          |    5 +++--
 conf/machine/fsg3be.conf                      |    2 +-
 conf/machine/kixrp435.conf                    |    2 +-
 conf/machine/nslu2be.conf                     |    2 +-
 conf/machine/nslu2le.conf                     |    2 +-
 recipes/binutils/binutils_csl-arm-20050416.bb |    2 +-
 recipes/binutils/binutils_csl-arm-20050603.bb |    2 +-
 recipes/eglibc/eglibc-package.bbclass         |    2 +-
 recipes/glibc/glibc-package.inc               |    2 +-
 recipes/prelink/prelink_20061027.bb           |    2 +-
 recipes/prelink/prelink_20071009.bb           |    2 +-
 recipes/prelink/prelink_20090925.bb           |    2 +-
 16 files changed, 26 insertions(+), 35 deletions(-)
conf/bitbake.conf +3 -14
index 494e14e..c2359dc 100644
@@ -675,20 +675,9 @@ BUILDCFG_NEEDEDVARS ?= "TARGET_ARCH TARGET_OS"
 
 # Overrides are processed left to right, so the ones that are named later take precedence.
 # You generally want them to go from least to most specific.
-# 
-# This means that an envionment variable named '<foo>_arm' overrides an
-# environment variable '<foo>' (when ${TARGET_ARCH} is arm).
-# an environment variable '<foo>_ramses' overrides '<foo>' but doesn't override
-# '<foo>_arm' when ${MACHINE} is 'ramses'. 
-# If you use combination ie '<foo>_arm_ramses', then '<foo>_arm_ramses' will override 
-# '<foo>_arm' and then '<foo>' will be overriden with that value from '<foo>_arm'.
-# And finally '<foo>_local' overrides anything, but with lowest priority.
-#
-# This works for  functions as well, they are really just environment variables.
-# Default OVERRIDES to make compilation fail fast in case of build system misconfiguration.
-OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
-# Alternative OVERRIDES definition without "fail fast", usually only for native building and Scratchbox toolchains.
-#OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:pn-${PN}"
+MACHINE_OVERRIDES ?= ""
+OVERRIDES = "pn-${PN}:fail-fast:build-${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:\
+${DISTRO}:${@':'.join(d.getVar('MACHINE_OVERRIDES', True).split())}:${MACHINE}:local"
 
 ##################################################################
 # Include the rest of the config files.
conf/distro/include/arm-thumb.inc +3 -3
index 36edd4f..75ed978 100644
@@ -16,9 +16,9 @@ THUMB_INTERWORK ?= "no"
 #    arm system and vice versa.  It is strongly recommended that DISTROs not
 #    turn this off - the actual cost is very small.
 
-OVERRIDE_THUMB = "${@['', ':thumb'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
-OVERRIDE_INTERWORK = "${@['', ':thumb-interwork'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
-OVERRIDES_append_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
+OVERRIDE_THUMB = "${@['', 'thumb:'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
+OVERRIDE_INTERWORK = "${@['', 'thumb-interwork:'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
+OVERRIDES_prepend_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
 
 #    Compiler and linker options for application code and kernel code.  These
 #    options ensure that the compiler has the correct settings for the selected
conf/distro/micro.conf +3 -3
index 6d396b3..19b5e62 100644
@@ -69,10 +69,10 @@ require conf/distro/include/sane-toolchain.inc
 require conf/distro/include/arm-thumb.inc
 
 #############################################################################
-# OVERRIDES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:\
-${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # PREFERRED VERSIONS
conf/distro/minimal.conf +3 -2
index 376ffde..0d32cfe 100644
@@ -81,9 +81,10 @@ KERNEL = "kernel26"
 MACHINE_KERNEL_VERSION = "2.6"
 
 #############################################################################
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # TOOLCHAIN
conf/distro/shr.conf +3 -2
index ade8cba..74e3fe9 100644
@@ -151,9 +151,10 @@ KERNEL = "kernel26"
 MACHINE_KERNEL_VERSION = "2.6"
 
 #############################################################################
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # TOOLCHAIN
conf/machine/fsg3be.conf +1 -1
index a2ca95b..c5ddc25 100644
@@ -8,7 +8,7 @@ PACKAGE_EXTRA_ARCHS = "ixp4xxbe"
 
 MACHINE_ARCH = "ixp4xxbe"
 
-OVERRIDES = "local:${MACHINE}:ixp4xxbe:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx ixp4xxbe"
 
 # Match the vendor's latest kernel
 PREFERRED_PROVIDER_virtual/kernel ?= "fsg3-kernel"
conf/machine/kixrp435.conf +1 -1
index 8259127..b4a2fa6 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"
 
 PREFERRED_PROVIDER_virtual/kernel = "linux"
 PREFERRED_VERSION_linux = "2.6.30"
conf/machine/nslu2be.conf +1 -1
index 9bc92f1..88e6ed2 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "armeb"
 # Note: armv4b armv4tb armv5teb will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5eb ixp4xxbe"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"
 
 ROOT_FLASH_SIZE ?= "6"
 
conf/machine/nslu2le.conf +1 -1
index 62e47cb..edd0d94 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"
 
 ROOT_FLASH_SIZE ?= "6"
 
recipes/binutils/binutils_csl-arm-20050416.bb +1 -1
index ad45438..71803d1 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"
 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs"
 PV = "2.15.99+csl-arm+cvs20050416"
 PR = "r3"
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 DEFAULT_PREFERENCE = "-1"
 DEFAULT_PREFERENCE_arm-linux = "-1"
 DEFAULT_PREFERENCE_arm-linuxeabi = "-1"
recipes/binutils/binutils_csl-arm-20050603.bb +1 -1
index 595cb98..50e8677 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"
 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs"
 PV = "2.15.99+csl-arm+cvs20050603"
 PR = "r3"
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 DEFAULT_PREFERENCE = "-1"
 #DEFAULT_PREFERENCE_arm-linux = "1"
 #DEFAULT_PREFERENCE_arm-linuxeabi = "1"
recipes/eglibc/eglibc-package.bbclass +1 -1
index dbca1ab..c802bd5 100644
@@ -116,7 +116,7 @@ def get_eglibc_fpu_setting(bb, d):
 EXTRA_OECONF += "${@get_eglibc_fpu_setting(bb, d)}"
 EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"
 
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 
 do_configure_prepend() {
         sed -e "s#@BASH@#/bin/sh#" -i ${S}/elf/ldd.bash.in
recipes/glibc/glibc-package.inc +1 -1
index 0b263bb..8cd8564 100644
@@ -44,7 +44,7 @@ def get_glibc_fpu_setting(bb, d):
 EXTRA_OECONF += "${@get_glibc_fpu_setting(bb, d)}"
 EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"
 
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 
 do_install() {
 	oe_runmake install_root=${D} install
recipes/prelink/prelink_20061027.bb +1 -1
index ad2faf5..552117b 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 TARGET_OS_ORIG := "${TARGET_OS}"
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"
recipes/prelink/prelink_20071009.bb +1 -1
index faf930c..e2ab3fb 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 TARGET_OS_ORIG := "${TARGET_OS}"
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"
recipes/prelink/prelink_20090925.bb +0 -0
index 4c83402..5b247b6 100644

recipes/prelink/prelink_20090925.bb SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \ · line 18

Quoted 8 lines
recipes/prelink/prelink_20090925.bb +1 -1
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 #TARGET_OS_ORIG := "${TARGET_OS}"
-#OVERRIDES_append = ":${TARGET_OS_ORIG}"
+#OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 #SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFMuDAvMkyGM64RGpERAlxxAKCzjaugv4f3l8A/wL+evxo3QOjbyQCgiuLC GTiLjxOoQ/7pW0vhSZf1owM=
=hkIm
-----END PGP SIGNATURE-----

#5 lore raw
In reply to Chris Larson
Quoted 9 lines
conf/machine/kixrp435.conf +0 -0
index 8259127..b4a2fa6 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc

automatically.
PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

Quoted 4 lines

-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build -${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"

Should this be "ixp4xx nslu2"? At least in order to keep the original behavior?

Quoted 12 lines

PREFERRED_PROVIDER_virtual/kernel = "linux"
PREFERRED_VERSION_linux = "2.6.30"

conf/machine/nslu2be.conf +0 -0
index 9bc92f1..88e6ed2 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "armeb"
 # Note: armv4b armv4tb armv5teb will be added by tune-xscale.inc

automatically.
PACKAGE_EXTRA_ARCHS = "armv5eb ixp4xxbe"

Quoted 4 lines

-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build -${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"

Should this be "ixp4xx nslu2"? At least in order to keep the original behavior?

Quoted 12 lines

ROOT_FLASH_SIZE ?= "6"

conf/machine/nslu2le.conf +0 -0
index 62e47cb..edd0d94 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc

automatically.
PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

Quoted 4 lines

-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build -${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"

Should this be "ixp4xx nslu2"? At least in order to keep the original behavior?

Quoted 5 lines

_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

#6 lore raw
Quoted 9 lines · Chase
Quoted 9 lines
conf/machine/kixrp435.conf +0 -0
index 8259127..b4a2fa6 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc

automatically.
PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

Quoted 8 lines
Quoted 2 lines

-OVERRIDES =

"local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build

Quoted 2 lines

-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"

Should this be "ixp4xx nslu2"? At least in order to keep the original behavior?

Yes, good catch, thanks, I'll fix it in the next version of the patch.

Signature · 5 lines
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
#7 lore raw
Quoted 25 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Chris Larson
Sent: Friday, October 15, 2010 10:17 AM
To: [email protected]
Cc: Chris Larson
Subject: Re: [oe] [PATCH] Reverse the order of OVERRIDES

Quoted 9 lines · Chase
Quoted 9 lines
conf/machine/kixrp435.conf +0 -0
index 8259127..b4a2fa6 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc

automatically.
PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

Quoted 3 lines
Quoted 2 lines

-OVERRIDES =

"local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "nslu2 ixp4xx"

Should this be "ixp4xx nslu2"? At least in order to keep the original behavior?

Yes, good catch, thanks, I'll fix it in the next version of the patch.

Everything else looked OK to me. Thanks for this change Chris.

Quoted 10 lines

--
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

#8 lore raw

PATCH 1/1 Reverse the order of OVERRIDES

In reply to Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that. As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific. As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

Big thanks to Chase Maupin for investigating and identifying this long standing issue.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so this implements this reversal, and drops the unnecessary and confusing comments.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally.

---
 conf/bitbake.conf                             |   17 ++-----------
 conf/distro/include/arm-thumb.inc             |    6 ++--
 conf/distro/micro.conf                        |    6 ++--
 conf/distro/minimal.conf                      |    5 ++-
 conf/distro/shr.conf                          |    5 ++-
 conf/machine/fsg3be.conf                      |    2 +-
 conf/machine/kixrp435.conf                    |    2 +-
 conf/machine/nslu2be.conf                     |    2 +-
 conf/machine/nslu2le.conf                     |    2 +-
 docs/usermanual/chapters/common_use_cases.xml |   31 +++++++++++++++++++-----
 recipes/binutils/binutils_csl-arm-20050416.bb |    2 +-
 recipes/binutils/binutils_csl-arm-20050603.bb |    2 +-
 recipes/eglibc/eglibc-package.bbclass         |    2 +-
 recipes/glibc/glibc-package.inc               |    2 +-
 recipes/prelink/prelink_20061027.bb           |    2 +-
 recipes/prelink/prelink_20071009.bb           |    2 +-
 recipes/prelink/prelink_20090925.bb           |    2 +-
 17 files changed, 50 insertions(+), 42 deletions(-)
conf/bitbake.conf +3 -14
index a5d6a1a..30dcd3e 100644
@@ -675,20 +675,9 @@ BUILDCFG_NEEDEDVARS ?= "TARGET_ARCH TARGET_OS"
 
 # Overrides are processed left to right, so the ones that are named later take precedence.
 # You generally want them to go from least to most specific.
-# 
-# This means that an envionment variable named '<foo>_arm' overrides an
-# environment variable '<foo>' (when ${TARGET_ARCH} is arm).
-# an environment variable '<foo>_ramses' overrides '<foo>' but doesn't override
-# '<foo>_arm' when ${MACHINE} is 'ramses'. 
-# If you use combination ie '<foo>_arm_ramses', then '<foo>_arm_ramses' will override 
-# '<foo>_arm' and then '<foo>' will be overriden with that value from '<foo>_arm'.
-# And finally '<foo>_local' overrides anything, but with lowest priority.
-#
-# This works for  functions as well, they are really just environment variables.
-# Default OVERRIDES to make compilation fail fast in case of build system misconfiguration.
-OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
-# Alternative OVERRIDES definition without "fail fast", usually only for native building and Scratchbox toolchains.
-#OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:pn-${PN}"
+MACHINE_OVERRIDES ?= ""
+OVERRIDES = "pn-${PN}:fail-fast:build-${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:\
+${DISTRO}:${@':'.join(d.getVar('MACHINE_OVERRIDES', True).split())}:${MACHINE}:local"
 
 ##################################################################
 # Include the rest of the config files.
conf/distro/include/arm-thumb.inc +3 -3
index 36edd4f..75ed978 100644
@@ -16,9 +16,9 @@ THUMB_INTERWORK ?= "no"
 #    arm system and vice versa.  It is strongly recommended that DISTROs not
 #    turn this off - the actual cost is very small.
 
-OVERRIDE_THUMB = "${@['', ':thumb'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
-OVERRIDE_INTERWORK = "${@['', ':thumb-interwork'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
-OVERRIDES_append_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
+OVERRIDE_THUMB = "${@['', 'thumb:'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
+OVERRIDE_INTERWORK = "${@['', 'thumb-interwork:'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
+OVERRIDES_prepend_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
 
 #    Compiler and linker options for application code and kernel code.  These
 #    options ensure that the compiler has the correct settings for the selected
conf/distro/micro.conf +3 -3
index 6d396b3..19b5e62 100644
@@ -69,10 +69,10 @@ require conf/distro/include/sane-toolchain.inc
 require conf/distro/include/arm-thumb.inc
 
 #############################################################################
-# OVERRIDES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:\
-${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # PREFERRED VERSIONS
conf/distro/minimal.conf +3 -2
index 376ffde..0d32cfe 100644
@@ -81,9 +81,10 @@ KERNEL = "kernel26"
 MACHINE_KERNEL_VERSION = "2.6"
 
 #############################################################################
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # TOOLCHAIN
conf/distro/shr.conf +3 -2
index ade8cba..74e3fe9 100644
@@ -151,9 +151,10 @@ KERNEL = "kernel26"
 MACHINE_KERNEL_VERSION = "2.6"
 
 #############################################################################
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 #############################################################################
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"
 
 #############################################################################
 # TOOLCHAIN
conf/machine/fsg3be.conf +1 -1
index a2ca95b..c5ddc25 100644
@@ -8,7 +8,7 @@ PACKAGE_EXTRA_ARCHS = "ixp4xxbe"
 
 MACHINE_ARCH = "ixp4xxbe"
 
-OVERRIDES = "local:${MACHINE}:ixp4xxbe:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx ixp4xxbe"
 
 # Match the vendor's latest kernel
 PREFERRED_PROVIDER_virtual/kernel ?= "fsg3-kernel"
conf/machine/kixrp435.conf +1 -1
index 8259127..ba8216d 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx nslu2"
 
 PREFERRED_PROVIDER_virtual/kernel = "linux"
 PREFERRED_VERSION_linux = "2.6.30"
conf/machine/nslu2be.conf +1 -1
index 9bc92f1..69d12a7 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "armeb"
 # Note: armv4b armv4tb armv5teb will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5eb ixp4xxbe"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx nslu2"
 
 ROOT_FLASH_SIZE ?= "6"
 
conf/machine/nslu2le.conf +1 -1
index 62e47cb..5cd8a77 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.
 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"
 
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx nslu2"
 
 ROOT_FLASH_SIZE ?= "6"
 
docs/usermanual/chapters/common_use_cases.xml +24 -7
index 8202dda..2114047 100644
@@ -7,8 +7,8 @@
 
     <para>Creating a new distribution is not complicated, however we urge you
     to try existing distributions first, because it's also very easy to do
-    wrong. The config needs to be created in $OEBASE/openembedded/conf/distro 
-    directory. So what has to be inside? 
+    wrong. The config needs to be created in $OEBASE/openembedded/conf/distro
+    directory. So what has to be inside?
       <itemizedlist>
         <listitem>
           <para><command>DISTRO_VERSION</command> so users will know which
@@ -72,8 +72,8 @@ SRCDATE = "20061014"
   <section id="commonuse_new_machine">
     <title>Adding a new Machine</title>
 
-    <para>To be able to build for a device OpenEmbedded has to know about it, 
-    so a machine config file needs to be written. All of the machine 
+    <para>To be able to build for a device OpenEmbedded has to know about it,
+    so a machine config file needs to be written. All of the machine
     configs are stored in $OEBASE/openembedded/conf/machine/ directory.</para>
 
     <para>As usual some variables are required: <itemizedlist>
@@ -97,6 +97,23 @@ SRCDATE = "20061014"
     <para>There are also some optional variables that can be defined:
     <itemizedlist>
     <listitem>
+      <para>
+        <command>MACHINE_OVERRIDES</command> lists additional items to add to
+        the <command>OVERRIDES</command> variable, between the
+        <command>DISTRO</command> and the <command>MACHINE</command>.  This is
+        utilized to add overrides which are less specific than the machine,
+        but are nonetheless related to it, allowing us to define variables a
+        certain way for a group of machines, rather than for each individual
+        one.  As an example, this variable may be used by the distribution to
+        add <command>SOC_FAMILY</command> or <command>MACHINE_CLASS</command>.
+      </para>
+      <para>
+        Note that this variable is space separated, and should always be
+        manipulated with +=, to ensure it's built up incrementally, and no
+        additions are lost.
+      </para>
+    </listitem>
+    <listitem>
           <para><command>SOC_FAMILY</command> describes a family of processors
             that all share common features such as kernel versions,
             bootloaders, etc.  This is used to allow overrides for a whole
@@ -105,7 +122,7 @@ SRCDATE = "20061014"
             or local setting.
           </para>
           <para>
-            NOTE: SOC_FAMILY is different than MACHINE_CLASS in that 
+            NOTE: SOC_FAMILY is different than MACHINE_CLASS in that
             MACHINE_CLASS is intended to specify a grouping of devices
             that may have different processors but share common features.
             For example all OMAP3 devices can be described using the SOC_FAMILY
@@ -139,7 +156,7 @@ SRCDATE = "20061014"
             <listitem><para>for cvs: add 'PV = "1.1+cvs${SRCREV}"' to your bb file.</para></listitem>
           </itemizedlist>
         Accompany either with an entry to conf/distro/include/sane-srcrevs.inc for a revision that you know
-        builds successfully.  It is also common to define the stable SRCREV 
+        builds successfully.  It is also common to define the stable SRCREV
         for your package directly in the package recipe.
         </para>
         <para>
@@ -276,7 +293,7 @@ RDEPENDS_${PN} += "\
             <title>Putting it together</title>
             <para>In the previous two sections we have prepared the host and
             target side. One thing that is missing is combining the two newly
-            created tasks and actually creating the SDK. This is what we are 
+            created tasks and actually creating the SDK. This is what we are
             going to do now.</para>
 
             <para>Create <filename>meta-toolchain-YOU.bb</filename> in the
recipes/binutils/binutils_csl-arm-20050416.bb +1 -1
index ad45438..71803d1 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"
 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs"
 PV = "2.15.99+csl-arm+cvs20050416"
 PR = "r3"
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 DEFAULT_PREFERENCE = "-1"
 DEFAULT_PREFERENCE_arm-linux = "-1"
 DEFAULT_PREFERENCE_arm-linuxeabi = "-1"
recipes/binutils/binutils_csl-arm-20050603.bb +1 -1
index 595cb98..50e8677 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"
 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs"
 PV = "2.15.99+csl-arm+cvs20050603"
 PR = "r3"
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 DEFAULT_PREFERENCE = "-1"
 #DEFAULT_PREFERENCE_arm-linux = "1"
 #DEFAULT_PREFERENCE_arm-linuxeabi = "1"
recipes/eglibc/eglibc-package.bbclass +1 -1
index dbca1ab..c802bd5 100644
@@ -116,7 +116,7 @@ def get_eglibc_fpu_setting(bb, d):
 EXTRA_OECONF += "${@get_eglibc_fpu_setting(bb, d)}"
 EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"
 
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 
 do_configure_prepend() {
         sed -e "s#@BASH@#/bin/sh#" -i ${S}/elf/ldd.bash.in
recipes/glibc/glibc-package.inc +1 -1
index 0b263bb..8cd8564 100644
@@ -44,7 +44,7 @@ def get_glibc_fpu_setting(bb, d):
 EXTRA_OECONF += "${@get_glibc_fpu_setting(bb, d)}"
 EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"
 
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 
 do_install() {
 	oe_runmake install_root=${D} install
recipes/prelink/prelink_20061027.bb +1 -1
index ad2faf5..552117b 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 TARGET_OS_ORIG := "${TARGET_OS}"
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"
recipes/prelink/prelink_20071009.bb +1 -1
index faf930c..e2ab3fb 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 TARGET_OS_ORIG := "${TARGET_OS}"
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"
recipes/prelink/prelink_20090925.bb +1 -1
index 4c83402..5b247b6 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 #TARGET_OS_ORIG := "${TARGET_OS}"
-#OVERRIDES_append = ":${TARGET_OS_ORIG}"
+#OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 #SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 
 S = "${WORKDIR}/prelink-0.0.${PV}"
Signature · 1 line
1.7.2.3
Acked-by on v1 1/1
#9 lore raw
Quoted 39 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Chris Larson
Sent: Friday, October 15, 2010 10:41 AM
To: [email protected]
Subject: [oe] [PATCH (v2)] Reverse the order of OVERRIDES

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that. As one example, "local"
is supposed to always be the most specific override, yet currently it's the
least specific. As another example, currently the target architecture is seen
as more specific than the machine, which is also clearly wrong.

Big thanks to Chase Maupin for investigating and identifying this long standing issue.

It becomes clear that a reversal of the current value will bring us to a more
sane behavior, and avoids the need for the dual overrides hack mentioned in
the comments, so this implements this reversal, and drops the unnecessary and
confusing comments.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to
inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to make
it easier and more convenient to assemble the variable incrementally.

Quoted 2 lines
Quoted 439 lines
---
 conf/bitbake.conf                             |   17 ++-----------
 conf/distro/include/arm-thumb.inc             |    6 ++--
 conf/distro/micro.conf                        |    6 ++--
 conf/distro/minimal.conf                      |    5 ++-
 conf/distro/shr.conf                          |    5 ++-
 conf/machine/fsg3be.conf                      |    2 +-
 conf/machine/kixrp435.conf                    |    2 +-
 conf/machine/nslu2be.conf                     |    2 +-
 conf/machine/nslu2le.conf                     |    2 +-
 docs/usermanual/chapters/common_use_cases.xml |   31 +++++++++++++++++++-
----
 recipes/binutils/binutils_csl-arm-20050416.bb |    2 +-
 recipes/binutils/binutils_csl-arm-20050603.bb |    2 +-
 recipes/eglibc/eglibc-package.bbclass         |    2 +-
 recipes/glibc/glibc-package.inc               |    2 +-
 recipes/prelink/prelink_20061027.bb           |    2 +-
 recipes/prelink/prelink_20071009.bb           |    2 +-
 recipes/prelink/prelink_20090925.bb           |    2 +-
 17 files changed, 50 insertions(+), 42 deletions(-)
conf/bitbake.conf +0 -0
index a5d6a1a..30dcd3e 100644
@@ -675,20 +675,9 @@ BUILDCFG_NEEDEDVARS ?= "TARGET_ARCH TARGET_OS"
 
 # Overrides are processed left to right, so the ones that are named later

take precedence.
# You generally want them to go from least to most specific.

conf/bitbake.conf +0 -4
-#
-# This means that an envionment variable named '<foo>_arm' overrides an
-# environment variable '<foo>' (when ${TARGET_ARCH} is arm).
-# an environment variable '<foo>_ramses' overrides '<foo>' but doesn't

override

conf/bitbake.conf +0 -2
-# '<foo>_arm' when ${MACHINE} is 'ramses'.
-# If you use combination ie '<foo>_arm_ramses', then '<foo>_arm_ramses'

will override
-# '<foo>_arm' and then '<foo>' will be overriden with that value from '<foo>_arm'.

conf/bitbake.conf +0 -3
-# And finally '<foo>_local' overrides anything, but with lowest priority.
-#
-# This works for  functions as well, they are really just environment

variables.
-# Default OVERRIDES to make compilation fail fast in case of build system misconfiguration.
-OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build- ${BUILD_OS}:fail-fast:pn-${PN}"
-# Alternative OVERRIDES definition without "fail fast", usually only for native building and Scratchbox toolchains.
-#OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build- ${BUILD_OS}:pn-${PN}"

conf/bitbake.conf +2 -0
+MACHINE_OVERRIDES ?= ""
+OVERRIDES = "pn-${PN}:fail-fast:build-

${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:\
+${DISTRO}:${@':'.join(d.getVar('MACHINE_OVERRIDES', True).split())}:${MACHINE}:local"

##################################################################
# Include the rest of the config files.

conf/distro/include/arm- renamed from conf/distro/include/arm-thumb.inc +0 -0

thumb.inc
index 36edd4f..75ed978 100644

conf/distro/include/arm-thumb.inc +0 -0
@@ -16,9 +16,9 @@ THUMB_INTERWORK ?= "no"
 #    arm system and vice versa.  It is strongly recommended that DISTROs

not
# turn this off - the actual cost is very small.

-OVERRIDE_THUMB = "${@['', ':thumb'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
-OVERRIDE_INTERWORK = "${@['', ':thumb- interwork'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"

conf/distro/include/arm-thumb.inc +1 -1
-OVERRIDES_append_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
+OVERRIDE_THUMB = "${@['', 'thumb:'][bb.data.getVar('ARM_INSTRUCTION_SET',

d, 1) == 'thumb']}"
+OVERRIDE_INTERWORK = "${@['', 'thumb- interwork:'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}" +OVERRIDES_prepend_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"

# Compiler and linker options for application code and kernel code.
These
# options ensure that the compiler has the correct settings for the
selected

conf/distro/micro.conf +3 -3
index 6d396b3..19b5e62 100644
@@ -69,10 +69,10 @@ require conf/distro/include/sane-toolchain.inc
 require conf/distro/include/arm-thumb.inc
 
 
 ##########################################################################
 ###
-# OVERRIDES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 
 ##########################################################################
 ###
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:\
-${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

##########################################################################
###
# PREFERRED VERSIONS

conf/distro/minimal.conf +1 -2
index 376ffde..0d32cfe 100644
@@ -81,9 +81,10 @@ KERNEL = "kernel26"
 MACHINE_KERNEL_VERSION = "2.6"
 
 
 ##########################################################################
 ###
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 
 ##########################################################################
 ###
-OVERRIDES =

"local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:b uild-${BUILD_OS}:fail-fast:pn-${PN}"

conf/distro/minimal.conf +2 -0
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

##########################################################################
###
# TOOLCHAIN

conf/distro/shr.conf +1 -2
index ade8cba..74e3fe9 100644
@@ -151,9 +151,10 @@ KERNEL = "kernel26"
 MACHINE_KERNEL_VERSION = "2.6"
 
 
 ##########################################################################
 ###
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES
 
 ##########################################################################
 ###
-OVERRIDES =

"local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:b uild-${BUILD_OS}:fail-fast:pn-${PN}"

conf/distro/shr.conf +2 -0
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

##########################################################################
###
# TOOLCHAIN

conf/machine/fsg3be.conf +0 -1
index a2ca95b..c5ddc25 100644
@@ -8,7 +8,7 @@ PACKAGE_EXTRA_ARCHS = "ixp4xxbe"
 
 MACHINE_ARCH = "ixp4xxbe"
 
-OVERRIDES =

"local:${MACHINE}:ixp4xxbe:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:bu ild-${BUILD_OS}:fail-fast:pn-${PN}"

conf/machine/fsg3be.conf +1 -0
+MACHINE_OVERRIDES += "ixp4xx ixp4xxbe"
 
 # Match the vendor's latest kernel
 PREFERRED_PROVIDER_virtual/kernel ?= "fsg3-kernel"
conf/machine/kixrp435.conf +0 -0
index 8259127..ba8216d 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc

automatically.
PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build -${BUILD_OS}:fail-fast:pn-${PN}"

conf/machine/kixrp435.conf +1 -0
+MACHINE_OVERRIDES += "ixp4xx nslu2"
 
 PREFERRED_PROVIDER_virtual/kernel = "linux"
 PREFERRED_VERSION_linux = "2.6.30"
conf/machine/nslu2be.conf +0 -0
index 9bc92f1..69d12a7 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "armeb"
 # Note: armv4b armv4tb armv5teb will be added by tune-xscale.inc

automatically.
PACKAGE_EXTRA_ARCHS = "armv5eb ixp4xxbe"

-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build -${BUILD_OS}:fail-fast:pn-${PN}"

conf/machine/nslu2be.conf +1 -0
+MACHINE_OVERRIDES += "ixp4xx nslu2"
 
 ROOT_FLASH_SIZE ?= "6"
conf/machine/nslu2le.conf +0 -0
index 62e47cb..5cd8a77 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
 # Note: armv4 armv4t armv5te will be added by tune-xscale.inc

automatically.
PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build -${BUILD_OS}:fail-fast:pn-${PN}"

conf/machine/nslu2le.conf +1 -0
+MACHINE_OVERRIDES += "ixp4xx nslu2"
 
 ROOT_FLASH_SIZE ?= "6"
docs/usermanual/chapters/common_use_cases.xml +0 -0
index 8202dda..2114047 100644
@@ -7,8 +7,8 @@
 
     <para>Creating a new distribution is not complicated, however we urge

you
to try existing distributions first, because it's also very easy to
do
- wrong. The config needs to be created in
$OEBASE/openembedded/conf/distro

docs/usermanual/chapters/common_use_cases.xml +1 -1
-    directory. So what has to be inside?
+    wrong. The config needs to be created in

$OEBASE/openembedded/conf/distro
+ directory. So what has to be inside?
<itemizedlist>
<listitem>
<para><command>DISTRO_VERSION</command> so users will know
which

docs/usermanual/chapters/common_use_cases.xml +2 -2
@@ -72,8 +72,8 @@ SRCDATE = "20061014"
   <section id="commonuse_new_machine">
     <title>Adding a new Machine</title>
 
-    <para>To be able to build for a device OpenEmbedded has to know about
 it,
-    so a machine config file needs to be written. All of the machine
+    <para>To be able to build for a device OpenEmbedded has to know about
 it,
+    so a machine config file needs to be written. All of the machine
     configs are stored in $OEBASE/openembedded/conf/machine/

directory.</para>

<para>As usual some variables are required: <itemizedlist>

docs/usermanual/chapters/common_use_cases.xml +2 -0
@@ -97,6 +97,23 @@ SRCDATE = "20061014"
     <para>There are also some optional variables that can be defined:
     <itemizedlist>
     <listitem>
+      <para>
+        <command>MACHINE_OVERRIDES</command> lists additional items to

add to

docs/usermanual/chapters/common_use_cases.xml +2 -0
+        the <command>OVERRIDES</command> variable, between the
+        <command>DISTRO</command> and the <command>MACHINE</command>.

This is
+ utilized to add overrides which are less specific than the
machine,
+ but are nonetheless related to it, allowing us to define
variables a
+ certain way for a group of machines, rather than for each
individual
+ one. As an example, this variable may be used by the
distribution to
+ add <command>SOC_FAMILY</command> or
<command>MACHINE_CLASS</command>.

docs/usermanual/chapters/common_use_cases.xml +4 -0
+      </para>
+      <para>
+        Note that this variable is space separated, and should always be
+        manipulated with +=, to ensure it's built up incrementally, and

no

docs/usermanual/chapters/common_use_cases.xml +4 -0
+        additions are lost.
+      </para>
+    </listitem>
+    <listitem>
           <para><command>SOC_FAMILY</command> describes a family of

processors
that all share common features such as kernel versions, bootloaders, etc. This is used to allow overrides for a
whole

docs/usermanual/chapters/common_use_cases.xml +1 -1
@@ -105,7 +122,7 @@ SRCDATE = "20061014"
             or local setting.
           </para>
           <para>
-            NOTE: SOC_FAMILY is different than MACHINE_CLASS in that
+            NOTE: SOC_FAMILY is different than MACHINE_CLASS in that
             MACHINE_CLASS is intended to specify a grouping of devices
             that may have different processors but share common features.
             For example all OMAP3 devices can be described using the

SOC_FAMILY

docs/usermanual/chapters/common_use_cases.xml +0 -0
@@ -139,7 +156,7 @@ SRCDATE = "20061014"
             <listitem><para>for cvs: add 'PV = "1.1+cvs${SRCREV}"' to

your bb file.</para></listitem>
</itemizedlist>
Accompany either with an entry to conf/distro/include/sane-
srcrevs.inc for a revision that you know
- builds successfully. It is also common to define the stable
SRCREV
+ builds successfully. It is also common to define the stable
SRCREV

         for your package directly in the package recipe.
         </para>
         <para>
docs/usermanual/chapters/common_use_cases.xml +0 -0
@@ -276,7 +293,7 @@ RDEPENDS_${PN} += "\
             <title>Putting it together</title>
             <para>In the previous two sections we have prepared the host

and
target side. One thing that is missing is combining the two
newly
- created tasks and actually creating the SDK. This is what we
are
+ created tasks and actually creating the SDK. This is what we
are
going to do now.</para>

<para>Create <filename>meta-toolchain-YOU.bb</filename> in
the

recipes/binutils/binutils_csl-arm-20050416.bb +1 -1
index ad45438..71803d1 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"
 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs"
 PV = "2.15.99+csl-arm+cvs20050416"
 PR = "r3"
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 DEFAULT_PREFERENCE = "-1"
 DEFAULT_PREFERENCE_arm-linux = "-1"
 DEFAULT_PREFERENCE_arm-linuxeabi = "-1"
recipes/binutils/binutils_csl-arm-20050603.bb +1 -1
index 595cb98..50e8677 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"
 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs"
 PV = "2.15.99+csl-arm+cvs20050603"
 PR = "r3"
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 DEFAULT_PREFERENCE = "-1"
 #DEFAULT_PREFERENCE_arm-linux = "1"
 #DEFAULT_PREFERENCE_arm-linuxeabi = "1"
recipes/eglibc/eglibc-package.bbclass +1 -1
index dbca1ab..c802bd5 100644
@@ -116,7 +116,7 @@ def get_eglibc_fpu_setting(bb, d):
 EXTRA_OECONF += "${@get_eglibc_fpu_setting(bb, d)}"
 EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"
 
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 
 do_configure_prepend() {
         sed -e "s#@BASH@#/bin/sh#" -i ${S}/elf/ldd.bash.in
recipes/glibc/glibc- renamed from recipes/glibc/glibc-package.inc +0 -0

package.inc
index 0b263bb..8cd8564 100644

recipes/glibc/glibc-package.inc +1 -1
@@ -44,7 +44,7 @@ def get_glibc_fpu_setting(bb, d):
 EXTRA_OECONF += "${@get_glibc_fpu_setting(bb, d)}"
 EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"
 
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"
 
 do_install() {
      oe_runmake install_root=${D} install
recipes/prelink/prelink_20061027.bb +1 -1
index ad2faf5..552117b 100644
@@ -12,7 +12,7 @@ SRC_URI =
 "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 TARGET_OS_ORIG := "${TARGET_OS}"
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 

S = "${WORKDIR}/prelink-0.0.${PV}"

recipes/prelink/prelink_20071009.bb +1 -1
index faf930c..e2ab3fb 100644
@@ -12,7 +12,7 @@ SRC_URI =
 "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 TARGET_OS_ORIG := "${TARGET_OS}"
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 

S = "${WORKDIR}/prelink-0.0.${PV}"

recipes/prelink/prelink_20090925.bb +1 -1
index 4c83402..5b247b6 100644
@@ -12,7 +12,7 @@ SRC_URI =
 "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \
            file://prelink.default"
 
 #TARGET_OS_ORIG := "${TARGET_OS}"
-#OVERRIDES_append = ":${TARGET_OS_ORIG}"
+#OVERRIDES_prepend = "${TARGET_OS_ORIG}:"
 #SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"
 

S = "${WORKDIR}/prelink-0.0.${PV}"
--
1.7.2.3

_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

Acked-by on v1 1/1
#10 lore raw
In reply to Chris Larson
Quoted 17 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that.  As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific.  As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

Big thanks to Chase Maupin for investigating and identifying this long standing issue.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so this implements this reversal, and drops the unnecessary and confusing comments.

Quoted 8 lines

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go.  This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally.

Quoted 346 lines

---
 conf/bitbake.conf                             |   17 ++----------- conf/distro/include/arm-thumb.inc             |    6 ++-- conf/distro/micro.conf                        |    6 ++-- conf/distro/minimal.conf                      |    5 ++- conf/distro/shr.conf                          |    5 ++- conf/machine/fsg3be.conf                      |    2 +-

 conf/machine/kixrp435.conf                    |    2 +-
 conf/machine/nslu2be.conf                     |    2 +-
 conf/machine/nslu2le.conf                     |    2 +-
 docs/usermanual/chapters/common_use_cases.xml |   31 +++++++++++++++++++-----
 recipes/binutils/binutils_csl-arm-20050416.bb |    2 +-
 recipes/binutils/binutils_csl-arm-20050603.bb |    2 +-
 recipes/eglibc/eglibc-package.bbclass         |    2 +-
 recipes/glibc/glibc-package.inc               |    2 +-
 recipes/prelink/prelink_20061027.bb           |    2 +-
 recipes/prelink/prelink_20071009.bb           |    2 +-
 recipes/prelink/prelink_20090925.bb           |    2 +-
 17 files changed, 50 insertions(+), 42 deletions(-)
conf/bitbake.conf +3 -14
index a5d6a1a..30dcd3e 100644
@@ -675,20 +675,9 @@ BUILDCFG_NEEDEDVARS ?= "TARGET_ARCH TARGET_OS"
 
  # Overrides are processed left to right, so the ones that are named later take precedence.
  # You generally want them to go from least to most specific.
-#
-# This means that an envionment variable named '<foo>_arm' overrides an
-# environment variable '<foo>' (when ${TARGET_ARCH} is arm).
-# an environment variable '<foo>_ramses' overrides '<foo>' but doesn't override
-# '<foo>_arm' when ${MACHINE} is 'ramses'.
-# If you use combination ie '<foo>_arm_ramses', then '<foo>_arm_ramses' will override
-# '<foo>_arm' and then '<foo>' will be overriden with that value from '<foo>_arm'.
-# And finally '<foo>_local' overrides anything, but with lowest priority.
-#
-# This works for  functions as well, they are really just environment variables.
-# Default OVERRIDES to make compilation fail fast in case of build system misconfiguration.
-OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
-# Alternative OVERRIDES definition without "fail fast", usually only for native building and Scratchbox toolchains.
-#OVERRIDES = "local:${MACHINE}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:pn-${PN}"
+MACHINE_OVERRIDES ?= ""
+OVERRIDES = "pn-${PN}:fail-fast:build-${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:\
+${DISTRO}:${@':'.join(d.getVar('MACHINE_OVERRIDES', True).split())}:${MACHINE}:local"
 
  ##################################################################
  # Include the rest of the config files.
conf/distro/include/arm-thumb.inc +3 -3
index 36edd4f..75ed978 100644
@@ -16,9 +16,9 @@ THUMB_INTERWORK ?= "no"
  #    arm system and vice versa.  It is strongly recommended that DISTROs not
  #    turn this off - the actual cost is very small.
 
-OVERRIDE_THUMB = "${@['', ':thumb'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
-OVERRIDE_INTERWORK = "${@['', ':thumb-interwork'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
-OVERRIDES_append_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
+OVERRIDE_THUMB = "${@['', 'thumb:'][bb.data.getVar('ARM_INSTRUCTION_SET', d, 1) == 'thumb']}"
+OVERRIDE_INTERWORK = "${@['', 'thumb-interwork:'][bb.data.getVar('THUMB_INTERWORK', d, 1) == 'yes']}"
+OVERRIDES_prepend_arm =. "${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}"
 
  #    Compiler and linker options for application code and kernel code.  These
  #    options ensure that the compiler has the correct settings for the selected
conf/distro/micro.conf +0 -0
index 6d396b3..19b5e62 100644
@@ -69,10 +69,10 @@ require conf/distro/include/sane-toolchain.inc

 require conf/distro/include/arm-thumb.inc

 #############################################################################

conf/distro/micro.conf +1 -1
-# OVERRIDES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES

 #############################################################################

conf/distro/micro.conf +2 -2
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:\
-${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

 #############################################################################
 # PREFERRED VERSIONS

conf/distro/minimal.conf +0 -0
index 376ffde..0d32cfe 100644
@@ -81,9 +81,10 @@ KERNEL = "kernel26"

 MACHINE_KERNEL_VERSION = "2.6"

 #############################################################################

conf/distro/minimal.conf +1 -1
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES

 #############################################################################

conf/distro/minimal.conf +2 -1
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

 #############################################################################
 # TOOLCHAIN

conf/distro/shr.conf +0 -0
index ade8cba..74e3fe9 100644
@@ -151,9 +151,10 @@ KERNEL = "kernel26"

 MACHINE_KERNEL_VERSION = "2.6"

 #############################################################################

conf/distro/shr.conf +1 -1
-# OVERWRITES adjusted from bitbake.conf to feature the MACHINE_CLASS
+# Ensure MACHINE_CLASS is in OVERRIDES

 #############################################################################

conf/distro/shr.conf +2 -1
-OVERRIDES = "local:${MACHINE}:${MACHINE_CLASS}:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_CLASS ?= ""
+MACHINE_OVERRIDES += "${MACHINE_CLASS}"

 #############################################################################
 # TOOLCHAIN

conf/machine/fsg3be.conf +0 -0
index a2ca95b..c5ddc25 100644
@@ -8,7 +8,7 @@ PACKAGE_EXTRA_ARCHS = "ixp4xxbe"
 

 MACHINE_ARCH = "ixp4xxbe"

conf/machine/fsg3be.conf +1 -1
-OVERRIDES = "local:${MACHINE}:ixp4xxbe:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx ixp4xxbe"

 # Match the vendor's latest kernel PREFERRED_PROVIDER_virtual/kernel ?= "fsg3-kernel"

conf/machine/kixrp435.conf +0 -0
index 8259127..ba8216d 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
  # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.

 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

conf/machine/kixrp435.conf +1 -1
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx nslu2"

 PREFERRED_PROVIDER_virtual/kernel = "linux"
 PREFERRED_VERSION_linux = "2.6.30"

conf/machine/nslu2be.conf +0 -0
index 9bc92f1..69d12a7 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "armeb"
  # Note: armv4b armv4tb armv5teb will be added by tune-xscale.inc automatically.

 PACKAGE_EXTRA_ARCHS = "armv5eb ixp4xxbe"

conf/machine/nslu2be.conf +1 -1
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx nslu2"

 ROOT_FLASH_SIZE ?= "6"

conf/machine/nslu2le.conf +0 -0
index 62e47cb..5cd8a77 100644
@@ -7,7 +7,7 @@ TARGET_ARCH = "arm"
  # Note: armv4 armv4t armv5te will be added by tune-xscale.inc automatically.

 PACKAGE_EXTRA_ARCHS = "armv5e ixp4xxle"

conf/machine/nslu2le.conf +1 -1
-OVERRIDES = "local:${MACHINE}:nslu2:ixp4xx:${DISTRO}:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}"
+MACHINE_OVERRIDES += "ixp4xx nslu2"

 ROOT_FLASH_SIZE ?= "6"

docs/usermanual/chapters/common_use_cases.xml +0 -0
index 8202dda..2114047 100644
@@ -7,8 +7,8 @@
 

    <para>Creating a new distribution is not complicated, however we urge you to try existing distributions first, because it's also very easy to do
-    wrong. The config needs to be created in $OEBASE/openembedded/conf/distro
-    directory. So what has to be inside?
+    wrong. The config needs to be created in $OEBASE/openembedded/conf/distro
+    directory. So what has to be inside?
      <itemizedlist>
        <listitem>
          <para><command>DISTRO_VERSION</command> so users will know which

docs/usermanual/chapters/common_use_cases.xml +0 -0
@@ -72,8 +72,8 @@ SRCDATE = "20061014"

  <section id="commonuse_new_machine">
    <title>Adding a new Machine</title>

-    <para>To be able to build for a device OpenEmbedded has to know about it,
-    so a machine config file needs to be written. All of the machine
+    <para>To be able to build for a device OpenEmbedded has to know about it,
+    so a machine config file needs to be written. All of the machine configs are stored in $OEBASE/openembedded/conf/machine/ directory.</para>

    <para>As usual some variables are required: <itemizedlist>

docs/usermanual/chapters/common_use_cases.xml +0 -0
@@ -97,6 +97,23 @@ SRCDATE = "20061014"

    <para>There are also some optional variables that can be defined:
    <itemizedlist>
    <listitem>
+      <para>
+        <command>MACHINE_OVERRIDES</command> lists additional items to add to
+        the <command>OVERRIDES</command> variable, between the
+        <command>DISTRO</command> and the <command>MACHINE</command>.  This is
+        utilized to add overrides which are less specific than the machine,
+        but are nonetheless related to it, allowing us to define variables a
+        certain way for a group of machines, rather than for each individual
+        one.  As an example, this variable may be used by the distribution to
+        add <command>SOC_FAMILY</command> or <command>MACHINE_CLASS</command>.
+      </para>
+      <para>
+        Note that this variable is space separated, and should always be
+        manipulated with +=, to ensure it's built up incrementally, and no
+        additions are lost.
+      </para>
+    </listitem>
+    <listitem>
          <para><command>SOC_FAMILY</command> describes a family of processors that all share common features such as kernel versions, bootloaders, etc.  This is used to allow overrides for a whole

docs/usermanual/chapters/common_use_cases.xml +0 -0
@@ -105,7 +122,7 @@ SRCDATE = "20061014"

            or local setting.
          </para>
          <para>
-            NOTE: SOC_FAMILY is different than MACHINE_CLASS in that
+            NOTE: SOC_FAMILY is different than MACHINE_CLASS in that MACHINE_CLASS is intended to specify a grouping of devices that may have different processors but share common features. For example all OMAP3 devices can be described using the SOC_FAMILY

docs/usermanual/chapters/common_use_cases.xml +0 -0
@@ -139,7 +156,7 @@ SRCDATE = "20061014"

            <listitem><para>for cvs: add 'PV = "1.1+cvs${SRCREV}"' to your bb file.</para></listitem> </itemizedlist>
        Accompany either with an entry to conf/distro/include/sane-srcrevs.inc for a revision that you know
-        builds successfully.  It is also common to define the stable SRCREV
+        builds successfully.  It is also common to define the stable SRCREV
        for your package directly in the package recipe.
        </para>
        <para>

docs/usermanual/chapters/common_use_cases.xml +0 -0
@@ -276,7 +293,7 @@ RDEPENDS_${PN} += "\

            <title>Putting it together</title>
            <para>In the previous two sections we have prepared the host and target side. One thing that is missing is combining the two newly
-            created tasks and actually creating the SDK. This is what we are
+            created tasks and actually creating the SDK. This is what we are going to do now.</para>

            <para>Create <filename>meta-toolchain-YOU.bb</filename> in the

recipes/binutils/binutils_csl-arm-20050416.bb +0 -0
index ad45438..71803d1 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"

 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs" PV = "2.15.99+csl-arm+cvs20050416"
 PR = "r3"

recipes/binutils/binutils_csl-arm-20050416.bb +1 -1
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"

 DEFAULT_PREFERENCE = "-1"
 DEFAULT_PREFERENCE_arm-linux = "-1"
 DEFAULT_PREFERENCE_arm-linuxeabi = "-1"

recipes/binutils/binutils_csl-arm-20050603.bb +0 -0
index 595cb98..50e8677 100644
@@ -7,7 +7,7 @@ LICENSE = "GPL"

 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/binutils-cvs" PV = "2.15.99+csl-arm+cvs20050603"
 PR = "r3"

recipes/binutils/binutils_csl-arm-20050603.bb +1 -1
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"

 DEFAULT_PREFERENCE = "-1"
 #DEFAULT_PREFERENCE_arm-linux = "1"
 #DEFAULT_PREFERENCE_arm-linuxeabi = "1"

recipes/eglibc/eglibc-package.bbclass +0 -0
index dbca1ab..c802bd5 100644
@@ -116,7 +116,7 @@ def get_eglibc_fpu_setting(bb, d):

 EXTRA_OECONF += "${@get_eglibc_fpu_setting(bb, d)}" EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"

recipes/eglibc/eglibc-package.bbclass +1 -1
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"

 do_configure_prepend() {
        sed -e "s#@BASH@#/bin/sh#" -i ${S}/elf/ldd.bash.in

recipes/glibc/glibc-package.inc +0 -0
index 0b263bb..8cd8564 100644
@@ -44,7 +44,7 @@ def get_glibc_fpu_setting(bb, d):

 EXTRA_OECONF += "${@get_glibc_fpu_setting(bb, d)}" EXTRA_OEMAKE += "rootsbindir=${base_sbindir}"

recipes/glibc/glibc-package.inc +1 -1
-OVERRIDES_append = ":${TARGET_ARCH}-${TARGET_OS}"
+OVERRIDES_prepend = "${TARGET_ARCH}-${TARGET_OS}:"

 do_install() {
       oe_runmake install_root=${D} install

recipes/prelink/prelink_20061027.bb +0 -0
index ad2faf5..552117b 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \

           file://prelink.default"

 TARGET_OS_ORIG := "${TARGET_OS}"

recipes/prelink/prelink_20061027.bb +1 -1
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"

 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"

 S = "${WORKDIR}/prelink-0.0.${PV}"

recipes/prelink/prelink_20071009.bb +0 -0
index faf930c..e2ab3fb 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \

           file://prelink.default"

 TARGET_OS_ORIG := "${TARGET_OS}"

recipes/prelink/prelink_20071009.bb +1 -1
-OVERRIDES_append = ":${TARGET_OS_ORIG}"
+OVERRIDES_prepend = "${TARGET_OS_ORIG}:"

 SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"

 S = "${WORKDIR}/prelink-0.0.${PV}"

recipes/prelink/prelink_20090925.bb +0 -0
index 4c83402..5b247b6 100644
@@ -12,7 +12,7 @@ SRC_URI = "${DEBIAN_MIRROR}/main/p/prelink/prelink_0.0.${PV}.orig.tar.gz \

           file://prelink.default"

 #TARGET_OS_ORIG := "${TARGET_OS}"

recipes/prelink/prelink_20090925.bb +1 -1
-#OVERRIDES_append = ":${TARGET_OS_ORIG}"
+#OVERRIDES_prepend = "${TARGET_OS_ORIG}:"

 #SRC_URI_append_linux-gnueabi = " file://arm_eabi.patch"

 S = "${WORKDIR}/prelink-0.0.${PV}"
--
1.7.2.3

_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

Acked-by on v1 1/1
#11 lore raw
In reply to Chris Larson

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 17 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that. As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific. As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

Big thanks to Chase Maupin for investigating and identifying this long standing issue.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so this implements this reversal, and drops the unnecessary and confusing comments.

Quoted 8 lines

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFMuK2DMkyGM64RGpERAsrgAJ0QbM4VQaERuv72HzOGZJKGqGxljwCfUpHL wX5AXypHwuHP4cG9BKj7CjM=
=Jpyx
-----END PGP SIGNATURE-----

#12 lore raw
Quoted 34 lines · Koen Kooi

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 4 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most

specific,

Quoted 1 line

however, our current usage of it does not match that. As one example,

"local"

Quoted 1 line

is supposed to always be the most specific override, yet currently it's

the

Quoted 1 line

least specific. As another example, currently the target architecture is

seen

Quoted 6 lines

as more specific than the machine, which is also clearly wrong.

Big thanks to Chase Maupin for investigating and identifying this long standing issue.

It becomes clear that a reversal of the current value will bring us to a

more

Quoted 1 line

sane behavior, and avoids the need for the dual overrides hack mentioned

in

Quoted 1 line

the comments, so this implements this reversal, and drops the unnecessary

and

Quoted 3 lines

confusing comments.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism

to

Quoted 1 line

inject overrides elements which are more specific than the distro but

less

Quoted 2 lines

specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to

Quoted 8 lines

make

Quoted 4 lines

it easier and more convenient to assemble the variable incrementally.

This is now in master -- thanks to all for the acks, review, comments -- let me know if any problems result from this.

Signature · 5 lines
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
#13 lore raw
Quoted 39 lines · Chris Larson
Quoted 37 lines · Koen Kooi

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 4 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most

specific,

Quoted 1 line

however, our current usage of it does not match that. As one example,

"local"

Quoted 1 line

is supposed to always be the most specific override, yet currently it's

the

Quoted 1 line

least specific. As another example, currently the target architecture is

seen

Quoted 6 lines

as more specific than the machine, which is also clearly wrong.

Big thanks to Chase Maupin for investigating and identifying this long standing issue.

It becomes clear that a reversal of the current value will bring us to a

more

Quoted 1 line

sane behavior, and avoids the need for the dual overrides hack mentioned

in

Quoted 1 line

the comments, so this implements this reversal, and drops the unnecessary

and

Quoted 3 lines

confusing comments.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism

to

Quoted 1 line

inject overrides elements which are more specific than the distro but

less

Quoted 2 lines

specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to

make

Quoted 2 lines

it easier and more convenient to assemble the variable incrementally.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

This is now in master -- thanks to all for the acks, review, comments -- let me know if any problems result from this.

You do realise the damage this potentially causes for compatibility of metadata between OE and Poky?

This change is pretty serious and potentially alters the handling of any double override. Poky uses them a bit more extensively than OE does. Its effectively an architecture change to OE yet no discussion was had at any TSC meeting :(.

I even asked about this a while back and was *told* that "local" was meant to be weak, I therefore added a strong version to Poky, in the spirit of maintaining compatibility.

Richard

#14 lore raw

On Wed, Nov 10, 2010 at 12:26 PM, Richard Purdie <[email protected]> wrote:
...

Quoted 4 lines

This change is pretty serious and potentially alters the handling of any double override. Poky uses them a bit more extensively than OE does. Its effectively an architecture change to OE yet no discussion was had at any TSC meeting :(.

...

I agree with Richard here. It was merged too fast.

We are suppose to have a release soon and this kind of change shouldn't be done so near of it.

Signature · 3 lines
Otavio Salvador                  O.S. Systems
E-mail: [email protected]  http://www.ossystems.com.br
Mobile: +55 53 9981-7854         http://projetos.ossystems.com.br
#15 lore raw
Quoted 6 lines · Otavio Salvador

On Wed, Nov 10, 2010 at 12:26 PM, Richard Purdie <[email protected]> wrote:
...

Quoted 4 lines

This change is pretty serious and potentially alters the handling of any double override. Poky uses them a bit more extensively than OE does. Its effectively an architecture change to OE yet no discussion was had at any TSC meeting :(.

Quoted 6 lines

...

I agree with Richard here. It was merged too fast.

We are suppose to have a release soon and this kind of change shouldn't be done so near of it.

To be fair, this was done a few weeks before it was decided that we'd even have releases, if I follow the timeline for that decision right.

Signature · 2 lines
Tom Rini
Mentor Graphics Corporation
#16 lore raw
In reply to Richard Purdie
Quoted 50 lines · Richard Purdie
Quoted 37 lines · Chris Larson
Quoted 35 lines · Koen Kooi

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 4 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most

specific,

Quoted 1 line

however, our current usage of it does not match that. As one example,

"local"

Quoted 1 line

is supposed to always be the most specific override, yet currently it's

the

Quoted 1 line

least specific. As another example, currently the target architecture is

seen

Quoted 6 lines

as more specific than the machine, which is also clearly wrong.

Big thanks to Chase Maupin for investigating and identifying this long standing issue.

It becomes clear that a reversal of the current value will bring us to a

more

Quoted 1 line

sane behavior, and avoids the need for the dual overrides hack mentioned

in

Quoted 1 line

the comments, so this implements this reversal, and drops the unnecessary

and

Quoted 3 lines

confusing comments.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism

to

Quoted 1 line

inject overrides elements which are more specific than the distro but

less

Quoted 2 lines

specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go. This variable is *space* separated, to

make

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

it easier and more convenient to assemble the variable incrementally.

This is now in master -- thanks to all for the acks, review, comments -- let me know if any problems result from this.

You do realise the damage this potentially causes for compatibility of metadata between OE and Poky?

Quoted 9 lines

This change is pretty serious and potentially alters the handling of any double override. Poky uses them a bit more extensively than OE does. Its effectively an architecture change to OE yet no discussion was had at any TSC meeting :(.

I even asked about this a while back and was *told* that "local" was meant to be weak, I therefore added a strong version to Poky, in the spirit of maintaining compatibility.

(a) Eeep! and (b) That's pretty much the opposite of what the rest of the thread / discussion was, which is to say "local is supposed to be the final winner, why isn't it?"

Now... what do we do here?

Signature · 2 lines
Tom Rini
Mentor Graphics Corporation
#17 lore raw
Quoted 81 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Tom Rini
Sent: Wednesday, November 10, 2010 10:29 AM
To: [email protected]
Subject: Re: [oe] [PATCH (v2)] Reverse the order of OVERRIDES

Quoted 2 lines · Richard Purdie
Quoted 1 line · Chris Larson

On Fri, Oct 15, 2010 at 12:37 PM, Koen Kooi

Quoted 8 lines · <[email protected]>wrote:
Quoted 8 lines
Quoted 8 lines

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 3 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the

variable is

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

expected to contain elements in the order least specific to most

specific,

Quoted 1 line

however, our current usage of it does not match that. As one example,

"local"

Quoted 1 line

is supposed to always be the most specific override, yet currently

it's

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

the

Quoted 1 line

least specific. As another example, currently the target

architecture is

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

seen

Quoted 3 lines

as more specific than the machine, which is also clearly wrong.

Big thanks to Chase Maupin for investigating and identifying this

long

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

standing issue.

It becomes clear that a reversal of the current value will bring us

to a

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

more

Quoted 1 line

sane behavior, and avoids the need for the dual overrides hack

mentioned

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

in

Quoted 1 line

the comments, so this implements this reversal, and drops the

unnecessary

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

and

Quoted 3 lines

confusing comments.

This also introduces a MACHINE_OVERRIDES variable as a generic

mechanism

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

to

Quoted 1 line

inject overrides elements which are more specific than the distro but

less

Quoted 1 line

specific than the machine, which is where things like MACHINE_CLASS

or

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

SOC_FAMILY or the like would go. This variable is *space* separated,

to

Quoted 8 lines
Quoted 8 lines
Quoted 7 lines

make

Quoted 4 lines

it easier and more convenient to assemble the variable incrementally.

This is now in master -- thanks to all for the acks, review, comments -

- let

Quoted 5 lines
Quoted 1 line

me know if any problems result from this.

You do realise the damage this potentially causes for compatibility of metadata between OE and Poky?

Quoted 9 lines

This change is pretty serious and potentially alters the handling of any double override. Poky uses them a bit more extensively than OE does. Its effectively an architecture change to OE yet no discussion was had at any TSC meeting :(.

I even asked about this a while back and was *told* that "local" was meant to be weak, I therefore added a strong version to Poky, in the spirit of maintaining compatibility.

Quoted 5 lines

(a) Eeep! and (b) That's pretty much the opposite of what the rest of the thread / discussion was, which is to say "local is supposed to be the final winner, why isn't it?"

Now... what do we do here?

I know this was all posted a while back. Has there been any resolution on this? I didn't notice any additional responses to this thread.

Quoted 9 lines

--
Tom Rini
Mentor Graphics Corporation

_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

#18 lore raw
Quoted 2 lines · Chase

I know this was all posted a while back.  Has there been any resolution on this?  I didn't notice any additional responses to this thread.

See the TSC meeting notes, it was discussed (with myself excluded from the decision making), and the decision was that it would be reverted and the issue revisited with an eye toward compatibility issues on a branch.

Signature · 5 lines
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
#19 lore raw
In reply to Chris Larson
Quoted 31 lines · Chris Larson

From: Chris Larson <[email protected]>

Given the current implementation of OVERRIDES in bitbake, the variable is expected to contain elements in the order least specific to most specific, however, our current usage of it does not match that.  As one example, "local" is supposed to always be the most specific override, yet currently it's the least specific.  As another example, currently the target architecture is seen as more specific than the machine, which is also clearly wrong.

It becomes clear that a reversal of the current value will bring us to a more sane behavior, and avoids the need for the dual overrides hack mentioned in the comments, so clean those up as well.

This also introduces a MACHINE_OVERRIDES variable as a generic mechanism to inject overrides elements which are more specific than the distro but less specific than the machine, which is where things like MACHINE_CLASS or SOC_FAMILY or the like would go.  This variable is *space* separated, to make it easier and more convenient to assemble the variable incrementally, and it's then translated to : separated when used in OVERRIDES.

---
 conf/bitbake.conf                             |   17 +++-------------- conf/distro/include/arm-thumb.inc             |    6 +++--- conf/distro/micro.conf                        |    6 +++--- conf/distro/minimal.conf                      |    5 +++-- conf/distro/shr.conf                          |    5 +++-- conf/machine/fsg3be.conf                      |    2 +-

 conf/machine/kixrp435.conf                    |    2 +-
 conf/machine/nslu2be.conf                     |    2 +-
 conf/machine/nslu2le.conf                     |    2 +-
Quoted 8 lines

 recipes/binutils/binutils_csl-arm-20050416.bb |    2 +- recipes/binutils/binutils_csl-arm-20050603.bb |    2 +- recipes/eglibc/eglibc-package.bbclass         |    2 +- recipes/glibc/glibc-package.inc               |    2 +-

 recipes/prelink/prelink_20061027.bb           |    2 +-
 recipes/prelink/prelink_20071009.bb           |    2 +-
 recipes/prelink/prelink_20090925.bb           |    2 +-
 16 files changed, 26 insertions(+), 35 deletions(-)

I'm sorry to notice too late, but this changes breaks do_unpack behavior, machine dir in FILESPATHPKG are no longer preferred ie for xorg-conf unpack:

now I get:

NOTE: package xserver-xorg-conf-0.1-r36: task do_unpack: Started
NOTE: Unpacking xserver-xorg-conf/xorg.conf to /OE/tmpdir-shr/work/om-gta02-oe-linux-gnueabi/xserver-xorg-conf-0.1-r36/
NOTE: package xserver-xorg-conf-0.1-r36: task do_unpack: Succeeded

if I revert this commit:

NOTE: package xserver-xorg-conf-0.1-r36: task do_unpack: Started
NOTE: Unpacking xserver-xorg-conf/om-gta02/xorg.conf to /OE/tmpdir-shr/work/om-gta02-oe-linux-gnueabi/xserver-xorg-conf-0.1-r36/
NOTE: package xserver-xorg-conf-0.1-r36: task do_unpack: Succeeded

Works OK if there is no "default" file (ie after rm xserver-xorg-conf/xorg.conf).

Probably needs reversed OVERRIDES order also in bitbake.conf FILESPATH

FILESPATH = "${@':'.join([os.path.normpath(os.path.join(fp, p, o)) for fp in d.getVar('FILESPATHBASE', 1).split(':') for p in d.getVar('FILESPATHPKG', 1).split(':') for o in (d.getVar('OVERRIDES',
1) + ':').split(':')])}"

I'll try it and send patch.

Regards,

#20 lore raw
Quoted 7 lines · Martin Jansa

Probably needs reversed OVERRIDES order also in bitbake.conf FILESPATH

FILESPATH = "${@':'.join([os.path.normpath(os.path.join(fp, p, o)) for fp in d.getVar('FILESPATHBASE', 1).split(':') for p in d.getVar('FILESPATHPKG', 1).split(':') for o in (d.getVar('OVERRIDES',
1) + ':').split(':')])}"

Yes, you're right. OVERRIDES in bitbake is used such that the *last* defined overridden variable is used, but FILESPATH uses the *first* file it finds. Reversing it there should indeed fix it. Nicely spotted, thanks. I have some errands to run, so won't have a chance to apply a fix until this evening, so if someone else wants to get it in, feel free.

Signature · 5 lines
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
#21 lore raw
Quoted 10 lines · Chris Larson
Quoted 7 lines · Martin Jansa

Probably needs reversed OVERRIDES order also in bitbake.conf FILESPATH

FILESPATH = "${@':'.join([os.path.normpath(os.path.join(fp, p, o)) for fp in d.getVar('FILESPATHBASE', 1).split(':') for p in d.getVar('FILESPATHPKG', 1).split(':') for o in (d.getVar('OVERRIDES',
1) + ':').split(':')])}"

Quoted 5 lines

Yes, you're right.  OVERRIDES in bitbake is used such that the *last* defined overridden variable is used, but FILESPATH uses the *first* file it finds.  Reversing it there should indeed fix it.  Nicely spotted, thanks.  I have some errands to run, so won't have a chance to apply a fix until this evening, so if someone else wants to get it in, feel free.

It's a bit more complicated

OE om-gta02@shr ~/dev/recipes/xorg-xserver $ bitbake -e -b
xserver-xorg-conf_0.1.bb | grep ^OVERRIDES OVERRIDES="thumb:thumb-interwork:libc-glibc:pn-xserver-xorg-conf:fail-fast:build-linux:arm:linux-gnueabi:shr::om-gta02:local:armv4t:armv4t"

that empty :: between shr and om-gta02 (where MACHINE_OVERRIDES should be) is breaking it for both orders (ie ignoring all after om-gta02 or after shr from with reversed order)

This is part of filespath (for first FILESPATHPKG - ${PF}) from base_do_unpack in base.bbclass:
'/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/task-unpack' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/thumb' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/thumb-interwork' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/libc-glibc' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/pn-xserver-xorg-conf' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/fail-fast' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/build-linux' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/arm' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/linux-gnueabi' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/shr' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36'
-- here is the problem
'/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/om-gta02' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/local' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/armv4t' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/armv4t' '/OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36'

Regards,

#22 lore raw
Quoted 19 lines · Martin Jansa
Quoted 10 lines · Chris Larson
Quoted 7 lines · Martin Jansa

Probably needs reversed OVERRIDES order also in bitbake.conf FILESPATH

FILESPATH = "${@':'.join([os.path.normpath(os.path.join(fp, p, o)) for fp in d.getVar('FILESPATHBASE', 1).split(':') for p in d.getVar('FILESPATHPKG', 1).split(':') for o in (d.getVar('OVERRIDES',
1) + ':').split(':')])}"

Quoted 5 lines

Yes, you're right.  OVERRIDES in bitbake is used such that the *last* defined overridden variable is used, but FILESPATH uses the *first* file it finds.  Reversing it there should indeed fix it.  Nicely spotted, thanks.  I have some errands to run, so won't have a chance to apply a fix until this evening, so if someone else wants to get it in, feel free.

It's a bit more complicated

Quoted 7 lines

OE om-gta02@shr ~/dev/recipes/xorg-xserver $ bitbake -e -b
xserver-xorg-conf_0.1.bb | grep ^OVERRIDES OVERRIDES="thumb:thumb-interwork:libc-glibc:pn-xserver-xorg-conf:fail-fast:build-linux:arm:linux-gnueabi:shr::om-gta02:local:armv4t:armv4t"

that empty :: between shr and om-gta02 (where MACHINE_OVERRIDES should be) is breaking it for both orders (ie ignoring all after om-gta02 or after shr from with reversed order)

Even with those 2 patches I've just sent I still have wrong order of TARGET_ARCH OVERRIDES :/.

OE om-gta02@shr ~/dev/recipes/xorg-xserver $ bitbake -e -b xserver-xorg-conf_0.1.bb | grep ^FILESPATH= | sed "s/:/\n/g" | head -n 20
FILESPATH=" /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/local /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/om-gta02 /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/shr /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/linux-gnueabi /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/arm /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/build-linux /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/fail-fast /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/pn-xserver-xorg-conf /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/libc-glibc /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/thumb-interwork /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/thumb /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36 /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/local /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/om-gta02 /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/shr /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/linux-gnueabi
.....

any hint?

Regards,

#23 lore raw
Quoted 58 lines · Martin Jansa
Quoted 19 lines · Martin Jansa
Quoted 10 lines · Chris Larson
Quoted 7 lines · Martin Jansa

Probably needs reversed OVERRIDES order also in bitbake.conf FILESPATH

FILESPATH = "${@':'.join([os.path.normpath(os.path.join(fp, p, o)) for fp in d.getVar('FILESPATHBASE', 1).split(':') for p in d.getVar('FILESPATHPKG', 1).split(':') for o in (d.getVar('OVERRIDES',
1) + ':').split(':')])}"

Quoted 5 lines

Yes, you're right.  OVERRIDES in bitbake is used such that the *last* defined overridden variable is used, but FILESPATH uses the *first* file it finds.  Reversing it there should indeed fix it.  Nicely spotted, thanks.  I have some errands to run, so won't have a chance to apply a fix until this evening, so if someone else wants to get it in, feel free.

It's a bit more complicated

Quoted 7 lines

OE om-gta02@shr ~/dev/recipes/xorg-xserver $ bitbake -e -b
xserver-xorg-conf_0.1.bb | grep ^OVERRIDES OVERRIDES="thumb:thumb-interwork:libc-glibc:pn-xserver-xorg-conf:fail-fast:build-linux:arm:linux-gnueabi:shr::om-gta02:local:armv4t:armv4t"

that empty :: between shr and om-gta02 (where MACHINE_OVERRIDES should be) is breaking it for both orders (ie ignoring all after om-gta02 or after shr from with reversed order)

Even with those 2 patches I've just sent I still have wrong order of TARGET_ARCH OVERRIDES :/.

OE om-gta02@shr ~/dev/recipes/xorg-xserver $ bitbake -e -b xserver-xorg-conf_0.1.bb | grep ^FILESPATH= | sed "s/:/\n/g" | head -n 20
FILESPATH=" /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/local /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/om-gta02 /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/shr /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/linux-gnueabi /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/arm /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/build-linux /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/fail-fast /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/pn-xserver-xorg-conf /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/libc-glibc /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/thumb-interwork /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/thumb /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36 /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/local /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/om-gta02 /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/shr /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/linux-gnueabi
.....

Quoted 1 line

any hint?

Ah I have TARGET_ARCH (actually FEED_ARCH) twice in OVERRIDES, one from sane-toolchain.inc and second from sane-feed.inc (maybe we should drop one of them).

I have patch which preppends FEED_ARCH instead of appending it -OVERRIDES .= ":${FEED_ARCH}"
+OVERRIDES =. "${FEED_ARCH}:"

in few distro configs

#       modified:   ../../conf/distro/chinook-compat.conf
#       modified:   ../../conf/distro/include/kaeilos-2010.inc
#       modified:   ../../conf/distro/include/kaeilos.inc
#       modified:   ../../conf/distro/include/sane-feed.inc
#       modified:   ../../conf/distro/include/sane-toolchain.inc
#       modified:   ../../conf/distro/maemo5-compat.conf

but now I'm quite confused if it's right fix, because with this reversed order of OVERRIDES it's not enough to change from append to preppend, because it also depends on order how bitbake.conf/sane-*.inc/distro.conf are included together :/.

So now I have

OE om-gta02@shr ~/dev/recipes/xorg-xserver $ bitbake -e -b
xserver-xorg-conf_0.1.bb | grep ^OVERRIDES OVERRIDES="thumb:thumb-interwork:libc-glibc:armv4t:armv4t:pn-xserver-xorg-conf:fail-fast:build-linux:arm:linux-gnueabi:shr:om-gta02:local"

OE om-gta02@shr ~/dev/recipes/xorg-xserver $ bitbake -e -b xserver-xorg-conf_0.1.bb | grep ^FILESPATH= | sed "s/:/\n/g" | head -n 20
FILESPATH=" /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/local /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/om-gta02 /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/shr /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/linux-gnueabi /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/arm /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/build-linux /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/fail-fast /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/pn-xserver-xorg-conf /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/armv4t /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/libc-glibc /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/thumb-interwork /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36/thumb /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1-r36 /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/local /OE/dev/recipes/xorg-xserver/xserver-xorg-conf-0.1/om-gta02

but it doesn't seem right again :/

Regards,

#24 lore raw

PATCH 1/1 bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

In reply to Chris Larson

* empty MACHINE_OVERRIDES adds :: to OVERRIDES (then in FILESPATH skips every directory after this)
* strip() is needed to remove space in case some distribution has ie MACHINE_OVERRIDES ?= ""
and then
MACHINE_OVERRIDES += ""
or
MACHINE_OVERRIDES ?= "${MACHINE_CLASS}", where MACHINE_CLASS is empty again

---
 conf/bitbake.conf |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)
conf/bitbake.conf +3 -2
index 155c593..5c45e1e 100644
@@ -676,8 +676,9 @@ BUILDCFG_NEEDEDVARS ?= "TARGET_ARCH TARGET_OS"
 # Overrides are processed left to right, so the ones that are named later take precedence.
 # You generally want them to go from least to most specific.
 MACHINE_OVERRIDES ?= ""
-OVERRIDES = "pn-${PN}:fail-fast:build-${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:\
-${DISTRO}:${@':'.join(d.getVar('MACHINE_OVERRIDES', True).split())}:${MACHINE}:local"
+OVERRIDES = "pn-${PN}:fail-fast:build-${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:${DISTRO}:\
+${@[':'.join(d.getVar('MACHINE_OVERRIDES', True).split())+':',''][d.getVar('MACHINE_OVERRIDES', True).strip() == '']}\
+${MACHINE}:local"
 
 ##################################################################
 # Include the rest of the config files.
Signature · 1 line
1.7.3.1
Acked-by on v1 1/1
#25 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 9 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Martin Jansa
Sent: Saturday, October 16, 2010 2:25 PM
To: [email protected]
Subject: [oe] [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 9 lines

* empty MACHINE_OVERRIDES adds :: to OVERRIDES (then in FILESPATH skips every directory after this)
* strip() is needed to remove space in case some distribution has ie MACHINE_OVERRIDES ?= ""
and then
MACHINE_OVERRIDES += ""
or
MACHINE_OVERRIDES ?= "${MACHINE_CLASS}", where MACHINE_CLASS is empty
again

I saw this issue in my overlay of having a "::" in my OVERRIDES which caused my parsing to fail in angstrom.inc. Applying this fix from Martin resolves the parsing issue. My OVERRIDES appear to be in the correct order with this change:

# OVERRIDES=${OVERRIDE_THUMB}${OVERRIDE_INTERWORK}libc-glibc:pn-${PN}:fail-fast:build-${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:${DISTRO}:${@[':'.join(d.getVar('MACHINE_OVERRIDES', True).split())+':',''][d.getVar('MACHINE_OVERRIDES', True).strip() == '']}${MACHINE}:local OVERRIDES="thumb-interwork:libc-glibc:pn-helloworld:fail-fast:build-linux:arm:linux-gnueabi:angstrom:armv7a:omap3:am37x-evm:local"

And my FILESPATH ordering is:

helloworld-1.0.0-r1/local
helloworld-1.0.0-r1/am37x-evm
helloworld-1.0.0-r1/omap3
helloworld-1.0.0-r1/armv7a
helloworld-1.0.0-r1/angstrom
helloworld-1.0.0-r1/linux-gnueabi
helloworld-1.0.0-r1/arm
helloworld-1.0.0-r1/build-linux
helloworld-1.0.0-r1/fail-fast
helloworld-1.0.0-r1/pn-helloworld
helloworld-1.0.0-r1/libc-glibc
helloworld-1.0.0-r1/thumb-interwork
helloworld-1.0.0-r1/
.
.
.

Quoted 34 lines
---
 conf/bitbake.conf |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)
conf/bitbake.conf +0 -0
index 155c593..5c45e1e 100644
@@ -676,8 +676,9 @@ BUILDCFG_NEEDEDVARS ?= "TARGET_ARCH TARGET_OS"
 # Overrides are processed left to right, so the ones that are named later

take precedence.
# You generally want them to go from least to most specific. MACHINE_OVERRIDES ?= ""
-OVERRIDES = "pn-${PN}:fail-fast:build- ${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:\
-${DISTRO}:${@':'.join(d.getVar('MACHINE_OVERRIDES', True).split())}:${MACHINE}:local"
+OVERRIDES = "pn-${PN}:fail-fast:build- ${BUILD_OS}:${TARGET_ARCH}:${TARGET_OS}:${DISTRO}:\
+${@[':'.join(d.getVar('MACHINE_OVERRIDES', True).split())+':',''][d.getVar('MACHINE_OVERRIDES', True).strip() == '']}\
+${MACHINE}:local"

##################################################################
# Include the rest of the config files.
--
1.7.3.1

_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

#26 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 19 lines · Chase
Quoted 9 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Martin Jansa
Sent: Saturday, October 16, 2010 2:25 PM
To: [email protected]
Subject: [oe] [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 9 lines

* empty MACHINE_OVERRIDES adds :: to OVERRIDES (then in FILESPATH skips every directory after this)
* strip() is needed to remove space in case some distribution has ie MACHINE_OVERRIDES ?= ""
and then
MACHINE_OVERRIDES += ""
or
MACHINE_OVERRIDES ?= "${MACHINE_CLASS}", where MACHINE_CLASS is empty
again

Quoted 4 lines

I saw this issue in my overlay of having a "::" in my OVERRIDES which caused my parsing to fail in angstrom.inc. Applying this fix from Martin resolves the parsing issue. My OVERRIDES appear to be in the correct order with this change:

Can you explain this? "caused my parsing to fail" is not very useful information. What exactly failed?

Signature · 5 lines
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
#27 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 36 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Chris Larson
Sent: Tuesday, October 19, 2010 2:09 PM
To: [email protected]
Subject: Re: [oe] [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 3 lines · Chase
Quoted 3 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf

Of

Quoted 7 lines
Quoted 7 lines

Martin Jansa
Sent: Saturday, October 16, 2010 2:25 PM
To: [email protected]
Subject: [oe] [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

* empty MACHINE_OVERRIDES adds :: to OVERRIDES (then in FILESPATH

skips

Quoted 7 lines
Quoted 7 lines

every directory after this)
* strip() is needed to remove space in case some distribution has ie MACHINE_OVERRIDES ?= ""
and then
MACHINE_OVERRIDES += ""
or
MACHINE_OVERRIDES ?= "${MACHINE_CLASS}", where MACHINE_CLASS is

empty

Quoted 4 lines
Quoted 1 line

again

I saw this issue in my overlay of having a "::" in my OVERRIDES which caused my parsing to fail in angstrom.inc. Applying this fix from

Quoted 8 lines

Martin

Quoted 1 line

resolves the parsing issue. My OVERRIDES appear to be in the correct

order

Quoted 1 line

with this change:

Can you explain this? "caused my parsing to fail" is not very useful information. What exactly failed?

Chris,

Sorry, I thought this was an understood issue. I have put the error output I got below. Basically, when I got to line 509 of data.py the variables had the following values:

vars = set(['__functions_', '_', '__functions__'])

var = '__functions__'

overrides = ['pn-angstrom', 'fail-fast', 'build-linux', 'arm', 'INVALID', 'angstrom-2008.1', '', 'am37x-evm', 'local']

o = ''

OVERRIDES = 'pn-angstrom:fail-fast:build-linux:arm:INVALID:angstrom-2008.1::am37x-evm:local'

The current override being evaluated was the '' one.

The line being evaluated in my overlay was USERDISTRO := "${DISTRO}" in my copy of angstrom.inc

The output from my build was:

a0271661@sdit-build01:~/projects/arago-multi-build$ bitbake helloworldTraceback (most recent call last):
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/bin/bitbake", line 143, in <module>
    main()
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/bin/bitbake", line 123, in main
    cooker.parseConfiguration()
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/cooker.py", line 68, in parseConfiguration
    self.parseConfigurationFile( os.path.join( "conf", "bitbake.conf" ) )
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/cooker.py", line 390, in parseConfigurationFile
    self.configuration.data = bb.parse.handle( afile, self.configuration.data )
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/__init__.py", line 75, in handle
    return h['handle'](fn, data, include)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 167, in handle
    feeder(lineno, s, fn, data)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 215, in feeder
    include(fn, s, data, False)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 114, in include
    ret = handle(fn, data, True)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/__init__.py", line 75, in handle
    return h['handle'](fn, data, include)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 167, in handle
    feeder(lineno, s, fn, data)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 221, in feeder
    include(fn, s, data, "include required")
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 114, in include
    ret = handle(fn, data, True)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/__init__.py", line 75, in handle
    return h['handle'](fn, data, include)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/BBHandler.py", line 175, in handle
    feeder(lineno, s, fn, base_name, d)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/BBHandler.py", line 381, in feeder
    return ConfHandler.feeder(lineno, s, fn, d)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 192, in feeder
    bb.data.update_data(e)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/data.py", line 509, in update_data
    for var in vars:
RuntimeError: Set changed size during iteration
Quoted 10 lines

--
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

#28 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 22 lines · Chase
Quoted 9 lines
Quoted 2 lines

I saw this issue in my overlay of having a "::" in my OVERRIDES which caused my parsing to fail in angstrom.inc. Applying this fix from

Martin

Quoted 1 line

resolves the parsing issue. My OVERRIDES appear to be in the correct

order

Quoted 1 line

with this change:

Can you explain this? "caused my parsing to fail" is not very useful information. What exactly failed?

Chris,

Sorry, I thought this was an understood issue. I have put the error output I got below. Basically, when I got to line 509 of data.py the variables had the following values:

vars = set(['__functions_', '_', '__functions__'])

var = '__functions__'

overrides = ['pn-angstrom', 'fail-fast', 'build-linux', 'arm', 'INVALID', 'angstrom-2008.1', '', 'am37x-evm', 'local']

Quoted 8 lines

o = ''

OVERRIDES = 'pn-angstrom:fail-fast:build-linux:arm:INVALID:angstrom-2008.1::am37x-evm:local'

The current override being evaluated was the '' one.

The line being evaluated in my overlay was USERDISTRO := "${DISTRO}" in my copy of angstrom.inc

Chris,

FYI, the below "RuntimeError: Set changed size during iteration" exception happens in bitbake 1.8.18 but not the 1.10/mainline.

Signature · 1 line
Denys
Quoted 38 lines

The output from my build was:

a0271661@sdit-build01:~/projects/arago-multi-build$ bitbake helloworldTraceback (most recent call last):
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/bin/bitbake", line 143, in <module>
    main()
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/bin/bitbake", line 123, in main
    cooker.parseConfiguration()
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/cooker.py", line 68, in parseConfiguration
    self.parseConfigurationFile( os.path.join( "conf", "bitbake.conf" ) )
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/cooker.py", line 390, in parseConfigurationFile
    self.configuration.data = bb.parse.handle( afile, self.configuration.data )
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/__init__.py", line 75, in handle
    return h['handle'](fn, data, include)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 167, in handle
    feeder(lineno, s, fn, data)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 215, in feeder
    include(fn, s, data, False)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 114, in include
    ret = handle(fn, data, True)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/__init__.py", line 75, in handle
    return h['handle'](fn, data, include)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 167, in handle
    feeder(lineno, s, fn, data)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 221, in feeder
    include(fn, s, data, "include required")
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 114, in include
    ret = handle(fn, data, True)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/__init__.py", line 75, in handle
    return h['handle'](fn, data, include)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/BBHandler.py", line 175, in handle
    feeder(lineno, s, fn, base_name, d)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/BBHandler.py", line 381, in feeder
    return ConfHandler.feeder(lineno, s, fn, d)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/parse/parse_py/ConfHandler.py", line 192, in feeder
    bb.data.update_data(e)
  File "/home/a0271661/projects/arago-multi-build/arago-bitbake/lib/bb/data.py", line 509, in update_data
    for var in vars:
RuntimeError: Set changed size during iteration
#29 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 29 lines · Denys Dmytriyenko
Quoted 28 lines · Chase
Quoted 9 lines
Quoted 2 lines

I saw this issue in my overlay of having a "::" in my OVERRIDES which caused my parsing to fail in angstrom.inc. Applying this fix from

Martin

Quoted 1 line

resolves the parsing issue. My OVERRIDES appear to be in the correct

order

Quoted 1 line

with this change:

Can you explain this? "caused my parsing to fail" is not very useful information. What exactly failed?

Chris,

Sorry, I thought this was an understood issue. I have put the error output I got below. Basically, when I got to line 509 of data.py the variables had the following values:

vars = set(['__functions_', '_', '__functions__'])

var = '__functions__'

overrides = ['pn-angstrom', 'fail-fast', 'build-linux', 'arm', 'INVALID', 'angstrom-2008.1', '', 'am37x-evm', 'local']

o = ''

OVERRIDES = 'pn-angstrom:fail-fast:build-linux:arm:INVALID:angstrom-2008.1::am37x-evm:local'

The current override being evaluated was the '' one.

Quoted 7 lines
Quoted 2 lines

The line being evaluated in my overlay was USERDISTRO := "${DISTRO}" in my copy of angstrom.inc

Chris,

FYI, the below "RuntimeError: Set changed size during iteration" exception happens in bitbake 1.8.18 but not the 1.10/mainline.

Chris, Martin,

After more debugging, it seems the problem is caused by using bitbake 1.8.18 and conf/collections.inc at the same time, as that version of bitbake chokes on empty override "::", when it gets re-spawned from collections.inc. If I replace the use of COLLECTIONS with direct BBFILES/BBFILE_* assignments, bitbake parses everything just fine. It is reproducible right now on the latest OE mainline with bitbake 1.8.18 by just adding "require conf/collections.inc" in local.conf or bitbake.conf.

I see that Martin's patch was marked as "Superseded" and archived - is there a new patch coming? Or are there any side-effects of this patch? I would definitely like to push this patch ASAP to address our immediate problem. Thanks.

Signature · 1 line
Denys
#30 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 5 lines · Denys Dmytriyenko

I see that Martin's patch was marked as "Superseded" and archived - is there a
new patch coming? Or are there any side-effects of this patch? I would definitely like to push this patch ASAP to address our immediate problem.

The original problem with empty overrides was the FILESPATH issue, which was fixed by the commit that sanitized FILESPATH, hence the superseded. Clearly there's another issue with them, so this patch is necessary. Feel free to apply it with modification and push it, we can always drop it when we bump the minimum bitbake version to 1.10.x.

Signature · 5 lines
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
#31 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 8 lines · Chris Larson
Quoted 5 lines · Denys Dmytriyenko

I see that Martin's patch was marked as "Superseded" and archived - is there a
new patch coming? Or are there any side-effects of this patch? I would definitely like to push this patch ASAP to address our immediate problem.

Quoted 5 lines

The original problem with empty overrides was the FILESPATH issue, which was fixed by the commit that sanitized FILESPATH, hence the superseded. Clearly there's another issue with them, so this patch is necessary. Feel free to apply it with modification and push it, we can always drop it when we bump the minimum bitbake version to 1.10.x.

Thanks. I will push this patch shortly - I actually have it in my local repository applied, but I'm now facing another issue and trying to solve it first...

Signature · 1 line
Denys
#32 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

In reply to Denys Dmytriyenko
Quoted 48 lines · Denys Dmytriyenko
Quoted 29 lines · Denys Dmytriyenko
Quoted 28 lines · Chase
Quoted 9 lines
Quoted 2 lines

I saw this issue in my overlay of having a "::" in my OVERRIDES which caused my parsing to fail in angstrom.inc. Applying this fix from

Martin

Quoted 1 line

resolves the parsing issue. My OVERRIDES appear to be in the correct

order

Quoted 1 line

with this change:

Can you explain this? "caused my parsing to fail" is not very useful information. What exactly failed?

Chris,

Sorry, I thought this was an understood issue. I have put the error output I got below. Basically, when I got to line 509 of data.py the variables had the following values:

vars = set(['__functions_', '_', '__functions__'])

var = '__functions__'

overrides = ['pn-angstrom', 'fail-fast', 'build-linux', 'arm', 'INVALID', 'angstrom-2008.1', '', 'am37x-evm', 'local']

o = ''

OVERRIDES = 'pn-angstrom:fail-fast:build-linux:arm:INVALID:angstrom-2008.1::am37x-evm:local'

The current override being evaluated was the '' one.

Quoted 7 lines
Quoted 2 lines

The line being evaluated in my overlay was USERDISTRO := "${DISTRO}" in my copy of angstrom.inc

Chris,

FYI, the below "RuntimeError: Set changed size during iteration" exception happens in bitbake 1.8.18 but not the 1.10/mainline.

Chris, Martin,

After more debugging, it seems the problem is caused by using bitbake 1.8.18 and conf/collections.inc at the same time, as that version of bitbake chokes on empty override "::", when it gets re-spawned from collections.inc. If I replace the use of COLLECTIONS with direct BBFILES/BBFILE_* assignments, bitbake parses everything just fine. It is reproducible right now on the latest OE mainline with bitbake 1.8.18 by just adding "require conf/collections.inc" in local.conf or bitbake.conf.

Quoted 3 lines

I see that Martin's patch was marked as "Superseded" and archived - is there a new patch coming? Or are there any side-effects of this patch? I would definitely like to push this patch ASAP to address our immediate problem.

I've marked it as "Superseded" when Chris pushed that filtering and I've checked that it fixes my case.

Maybe we can resolve it by using limited OVERRIDES in FILESPATH as suggested in "[RFC] do we really need all OVERRIDES in FILESPATH?". Instead of creating more complicated constructs to be able to push more stuff to FILESPATH (without '::') and then never use it.

Regards,

Signature · 1 line
Martin 'JaMa' Jansa     jabber: [email protected]
#33 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 9 lines · Martin Jansa

I've marked it as "Superseded" when Chris pushed that filtering and I've checked
that it fixes my case.

Maybe we can resolve it by using limited OVERRIDES in FILESPATH as suggested in "[RFC] do we really need all OVERRIDES in FILESPATH?". Instead of creating more complicated constructs to be able to push more stuff to FILESPATH (without '::') and then never use it.

MACHINE_OVERRIDES aren't specific to FILESPATH, as far as I know.

Signature · 5 lines
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
#34 lore raw

Re: [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

In reply to Chris Larson
Quoted 36 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Chris Larson
Sent: Tuesday, October 19, 2010 2:09 PM
To: [email protected]
Subject: Re: [oe] [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

Quoted 3 lines · Chase
Quoted 3 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf

Of

Quoted 7 lines
Quoted 7 lines

Martin Jansa
Sent: Saturday, October 16, 2010 2:25 PM
To: [email protected]
Subject: [oe] [PATCH] bitbake.conf: don't append MACHINE_OVERRIDES to OVERRIDES when it's empty

* empty MACHINE_OVERRIDES adds :: to OVERRIDES (then in FILESPATH

skips

Quoted 7 lines
Quoted 7 lines

every directory after this)
* strip() is needed to remove space in case some distribution has ie MACHINE_OVERRIDES ?= ""
and then
MACHINE_OVERRIDES += ""
or
MACHINE_OVERRIDES ?= "${MACHINE_CLASS}", where MACHINE_CLASS is

empty

Quoted 4 lines
Quoted 1 line

again

I saw this issue in my overlay of having a "::" in my OVERRIDES which caused my parsing to fail in angstrom.inc. Applying this fix from

Quoted 8 lines

Martin

Quoted 1 line

resolves the parsing issue. My OVERRIDES appear to be in the correct

order

Quoted 1 line

with this change:

Can you explain this? "caused my parsing to fail" is not very useful information. What exactly failed?

One other detail I forgot to mention is that I was actually using bitbake from the Arago project which is based on 1.8.18. I did not see this issue if I used bitbake 1.10 I did not see this issue. I kept this same fix in my tree and tried with bitbake 1.10 and it too worked so it didn't seem like the fix was negatively impacting the 1.10 version.

Quoted 10 lines

--
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics
_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

#35 lore raw

PATCH 1/1 bitbake.conf: reverse OVERRIDES order in FILESPATH definition

In reply to Chris Larson

* directories in FILESPATH are search for first fit and we would expect ie ${PN}/local/a.patch to be used before ${PN}/ramses/a.patch
* moving last ':' before OVERRIDES makes FILESPATHBASE/FILESPATHPKG/ as last option instead of first

---
 conf/bitbake.conf |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
conf/bitbake.conf +1 -1
index edb9530..1a08673 100644
@@ -271,7 +271,7 @@ FILES_${PN}-locale = "${datadir}/locale"
 FILE_DIRNAME = "${@os.path.dirname(bb.data.getVar('FILE', d))}"
 FILESPATHBASE = "${FILE_DIRNAME}"
 FILESPATHPKG = "${PF}:${P}:${PN}:${BP}:${BPN}:files:."
-FILESPATH = "${@':'.join([os.path.normpath(os.path.join(fp, p, o)) for fp in d.getVar('FILESPATHBASE', 1).split(':') for p in d.getVar('FILESPATHPKG', 1).split(':') for o in (d.getVar('OVERRIDES', 1) + ':').split(':')])}"
+FILESPATH = "${@':'.join([os.path.normpath(os.path.join(fp, p, o)) for fp in d.getVar('FILESPATHBASE', 1).split(':') for p in d.getVar('FILESPATHPKG', 1).split(':') for o in (':' + d.getVar('OVERRIDES', 1)).split(':')[::-1]])}"
 FILESDIR = "${@bb.which(d.getVar('FILESPATH', 1), '.')}"
 
 ##################################################################
Signature · 1 line
1.7.3.1